mTOR inhibition improves the formation of functional T cell memory following COVID-19 vaccination of kidney transplant recipients

mTOR inhibitors are associated with superior induction of humoral and cellular immunity to primary COVID-19 vaccination as compared to calcineurin inhibitors in a three-drug kidney transplant maintenance regimen

Healthy individuals and KTRs were prospectively recruited to receive two doses of BNT162b2 or ChAdOx1 as part of the REVAX trial (ACTRN12621000532808; fig. S1) and in accordance with Australian health guidelines (Fig. 1). KTRs were receiving a variety of immunosuppressive drug regimens (Table 1), and healthy (non-transplanted) controls of comparable age and sex were recruited from amongst patients’ cohabitants (Table 1). Patients and cohabitants were vaccinated concurrently, and peripheral blood samples were collected three weeks (median 22 days, IQR 22-25) after the second vaccine dose for analysis of SARS-CoV-2-specific immunity (Fig. 1).

• Download figure
• Open in new tab

Figure 1. Study design.

Observational Study: Unvaccinated, SARS-CoV-2 infection-naïve kidney transplant recipients (KTR; n = 76) and healthy controls (HC; n = 71) were administered two doses of ChAdOx1 or BNT162b2 vaccine as part of the REVAX trial (ACTRN12621000532808), and peripheral blood samples were collected 21 days post-second dose. A comprehensive analysis of the functional humoral and cellular SARS-CoV-2 Spike-specific immune response was performed. Analysis 1: Spike-specific immunity was compared between HCs and KTRs, and between KTRs receiving CNI-based standard-of-care (‘SOC’; n = 48) and mTORi-based (‘mTORi’; n = 15) three-drug regimens. Analysis 2: Based on the high IFNγ ELISpot response observed for the mTORi group in Analysis 1, a second analysis was performed comparing Spike-specific T cell responses in KTRs receiving an mTORi (n = 25, independent of immunosuppresion protocol) to that of healthy individuals, in order to test the hypothesis that mTORi use is associated with a boost in the antigen-specific T cell memory response to vaccination. mTORi users were stratified into those receiving rapamycin (n = 14) and everolimus (n = 11), and the phenotype and function of Spike-specific T cells in this subcohort assessed by AIM and ICS assays by flow cytometry. Randomized controlled trial: a randomized, controlled trial was performed across two academic transplant centers in Australia to measure the effect of immunosuppression modification with rapamycin on humoral and cellular immune response to COVID-19 booster vaccination. KTRs receiving SOC triple therapy were randomized to remain on existing treatment or switch to the intervention arm: cease mycophenolate mofetil, start rapamycin and reduce tacrolimus. Participants received a third COVID-19 vaccine dose (BNT162b2), and cellular and humoral immunity was assessed 4 weeks later. Animal study: BALB/c mice (n=7 per group) were vaccinated (intra-muscularly (IM)) with two doses of BNT162b2 COVID-19 vaccine (1 μg), boosted with BNT162b2 (1 μg, IM), or with Omicron-RBD DNA vaccine (50 μg, intra-dermally (ID)). Mice were started on rapamycin prior to the primary vaccination course (PRI), or prior to booster vaccination (BST) with the homologous or Omicron variant vaccine, and the effect of rapamycin on humoral and cellular immune responses were evaluated. Created with BioRender.com.

Humoral immune response to vaccination was measured by total anti-Spike IgG titer, anti-receptor-binding domain (RBD) IgG titer, and live virus neutralization as the gold-standard correlate of protection from infection and severe disease (24). Cellular immunity was assessed by interferon (IFN)-γ ELISpot in response to stimulation with overlapping peptides spanning the SARS-CoV-2 Spike protein.

SARS-CoV-2-specific immunity was first compared between the complete KTR cohort and healthy controls (Fig. 2A, C, E). Consistent with previous reports, KTRs demonstrated significant impairment in anti-Spike IgG titer (median: AUC 1525 vs 0.8043, p<0.0001), seroconversion of anti-RBD IgG (100% vs 38.0%, p=0.0005), live virus neutralization (100% vs 18.3%, p<0.0001), and IFNγ-secreting Spike-specific T cell response (median: 430 vs 88 SFU/10⁶ cells, p<0.0001), irrespective of the vaccine type (ChAdOx or BNT162b2) (fig. S2).

• Download figure
• Open in new tab

Fig. 2. Immune response of kidney transplant recipients (KTRs) to a primary COVID-19 vaccination schedule.

Responses of KTRs (n = 76) are compared against healthy controls (HC; n = 71), and between clinically comparable calcineurin inhibitor-based (‘SOC’; n = 43) and mTOR inhibitor-based (‘mTORi’; n = 15) three-drug immunosuppressive regimens. (A-B) End point titers are reported as area under the curve (AUC). Dashed lines represent detection threshold (mean AUC + 2 SD of negative controls). (C-D) Log-scale comparison of anti-RBD IgG. The detection range for the assay is 0.4 – 250 U/mL, as demarcated by dotted lines. Percentage of seropositive individuals (above 0.4 U/mL) is shown for each group. (E-F) Endpoint titers (minimum serum dilution) for 50% neutralization of live SARS-CoV-2 virus (ancestral strain) entry into target cells. (G-H) Log-scale comparison of Spike-specific IFNγ-secreting T cell responses measured as spot-forming units (SFU) by ELISpot. Comparisons between groups by Mann-Whitney U test (flat bars) or Fisher’s exact test (crooked bars): ns, non-significant; *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

We next asked whether immunosuppression protocol has a significant effect on vaccine response. KTRs receiving standard-of-care (SOC) maintenance with a calcineurin inhibitor, anti-metabolite, and steroid were compared with those maintained on an alternative mTORi-based three-drug regimen (mTORi, anti-metabolite, steroid) (15), to assess which of these comparable protocols is preferable for an effective immune response to vaccination.

KTRs receiving the mTORi-based protocol (n = 15) demonstrated a median Spike-specific T cell response of 520 SFU/10⁶ cells, 12-fold higher than that of the SOC group (n = 43; p<0.0001), and similar to that of healthy controls (Fig. 2H). To confirm and further characterize the difference in T cell response between treatments, we evaluated the function and memory phenotype of Spike-specific CD4⁺ and CD8⁺ T cells in the mTORi group (n = 15), compared against a sub-cohort of KTRs on CNI-based SOC therapy (n = 15) and HCs (n = 15). Consistent with the ELISpot analysis, SOC immunosuppression was associated with impairment in the phenotype and function of Spike-specific effector memory T cells relative to healthy controls (fig. S3). Conversely, KTRs receiving the mTORi regimen formed T cell memory that was functionally and phenotypically similar to healthy individuals (fig. S3), and was superior to SOC in frequencies of all major CD4⁺ and CD8⁺ T cell memory populations, as well as greater frequencies of cytokine producing CD4⁺ and CD8⁺ T cells and stronger cytotoxic recall (GZMB⁺PRFN⁺) responses (fig. S3) (see Supplementary Materials).

Anti-Spike IgG titer was also significantly greater in mTORi-treated KTRs than SOC-treated (AUC 6.766 vs 0.4116, p=0.0011) (Fig. 2B). Despite this quantitative advantage, the two groups were equivalent in proportion of seroconverted patients with detectable anti-RBD IgG (33% vs 33%, p>0.99) (Fig. 2D), and in capacity of sera to neutralize live SARS-CoV-2 virus (ancestral strain A.2.2) (20.0% vs 11.9%, p=0.3910) (Fig. 2F).

The higher anti-Spike IgG titer for mTORi relative to SOC was surprising given the important role of mTORC1 signaling in the formation of T follicular helper (T_FH) cells and germinal centers in mice (26). To investigate this, the antibody response in the sub-cohort was investigated in greater detail. In addition to the elevated anti-Spike IgG, we found a higher titer of anti-Spike IgM, and a trend towards greater anti-Spike IgA, for mTORi compared to SOC (fig. S4). Consistent with a role for mTORC1 in T_FH cell development, patients receiving mTORi demonstrated a reduced frequency of circulating (c)T_FH cells (defined as CD4⁺CXCR5⁺ T cells) versus SOC (fig. S4). However, a greater proportion of cT_FH cells in the mTORi group upregulated CD154 (CD40L) in response to stimulation with Spike peptides, suggesting that the reduced T_FH cell frequency may be partially offset by a greater proportion of antigen-specific T_FH cells capable of providing B cell help (fig. S3). No significant difference in live virus neutralization of Delta and Omicron BA.1 variants was observed between groups (fig. S4).

Taken together, usage of mTORi-based rather than SOC maintenance immunosuppression by KTRs was associated with superior induction of cellular and humoral immune responses to COVID-19 vaccination.

KTRs receiving rapamycin have greater Spike-specific CD4⁺ and CD8⁺ memory T cell frequencies than healthy individuals

We next investigated whether mTORC1 inhibition promotes the formation of antigen-specific CD8⁺ memory T cells in humans, as previously described in animal models (17, 18). In support of this, KTRs on the mTORi-based triple therapy (mTORi, MMF, prednisolone) exhibited a trend towards a greater median IFNγ-secreting T cell response than healthy controls following primary vaccination (520 vs 430 SFU/10⁶ cells, p=0.1751; Fig. 2). We therefore extended our analysis to include all KTRs in our cohort who were receiving an mTORi as a component of two- or three-drug maintenance regimens, stratified by use of rapamycin (n = 14) or everolimus (n = 11) (Table 2) for comparison with healthy individuals.

KTRs receiving rapamycin as a component of their immunosuppression regimen were found to have significantly greater frequencies of IFNγ-secreting, Spike-specific T cells than healthy individuals following primary COVID-19 vaccination (1020 vs 430 SFU/10⁶ cells; p=0.0395, Fig. 3A).

• Download figure
• Open in new tab

Fig. 3. Phenotype and function of memory T cells following primary COVID-19 vaccination in healthy controls (HCs; n = 15) and kidney transplant recipients (KTRs) receiving rapamycin-inclusive (RAP-incl.; n = 14) and everolimus-inclusive (EVR-incl.; n = 11) immunosuppressive regimens.

(A) Log-scale comparison of Spike-specific IFNγ-secreting T cell responses, measured as spot-forming units (SFU) by ELISpot. (B) T cells were gated as CD3⁺CD19 ⁻CD14⁻, and antigen-specific T cells defined by activation-induced marker (AIM) expression as CD4⁺CD134⁺CD137⁺ or CD8⁺CD69⁺CD137⁺. Within the AIM⁺ gates, four populations were defined by expression of CCR7 and CD45RA. (C-D) Percentage of AIM⁺ cells within the CD4⁺ (C) and CD8⁺ (D) T cell compartments. (E-F) Frequency of each AIM⁺ subpopulation as a percentage of the total CD4⁺ (E) or CD8⁺ (F) T cell populations. (G-H) Contribution of each subpopulation to the CD4⁺ (G) and CD8⁺ (H) AIM⁺ populations. Naïve: T_N, CCR7⁺CD45RA⁺; Central Memory: T_CM, CCR7⁺CD45RA⁻; Effector Memory: T_EM, CCR7⁻CD45RA⁻; Terminally Differentiated Effector Memory: T_EMRA, CCR7⁻ CD45RA⁺. (I-J) Log-scale comparison of the percentage of CD4⁺ (I) and CD8⁺ (J) T cells positive for intracellular granzyme B (GZMB⁺) and perforin (PRFN⁺) following stimulation with Spike peptides. (K-L) Log-scale comparison of frequency of Spike-specific CD4⁺CD154⁺ (K) and CD8⁺ (L) T cells single, double, or triple positive for any combination of IL-2, TNF and IFNγ cytokines, as a percentage of CD4⁺ and CD8⁺, respectively. (M-N) Frequency of CD4⁺CD154⁺ (M) and CD8⁺ (N) T cells upregulating at-least one cytokine in response to stimulation, as a percentage of CD4⁺ and CD8⁺, respectively. (O-P) IFNγ (O) and TNF (P) mean fluorescence intensity (MFI) of Spike-specific CD4⁺CD154⁺ and CD8⁺ T cells. (Q-R) Median cytokine expression profile (outer pie chart) and proportion of cytokine positive cells that are polyfunctional (inner pie chart) of CD4⁺CD154⁺ (Q) and CD8⁺ (R) T cells. Statistical analysis by Kruskal-Wallis with Dunn’s correction comparing each group to healthy controls: ns, non-significant; *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

No significant difference in T cell response was observed between the healthy and everolimus groups (Fig. 3A). It is unclear whether this reflects an intrinsic difference between the two mTOR inhibitors or the greater burden of co-immunosuppression in the everolimus group (Table 2).

To investigate the elevated T cell response associated with rapamycin, Spike-specific T cell frequencies were assessed independent of cytokine production by activation-induced marker (AIM) assay. For this analysis, isolated PBMCs were stimulated for 24h with a Spike-derived peptide array and Spike-specific T cells identified as CD4⁺CD134⁺CD137⁺ or CD8⁺CD69⁺CD137⁺ (25, 26) (fig. S5; Fig. 3B). KTRs receiving rapamycin exhibited a 3.3-fold higher median frequency of Spike-specific CD4⁺ T cells (1.60% vs 0.48%, p=0.0003), and a 3.1-fold higher frequency of CD8⁺ T cells (0.86% vs 0.28%, p=0.0017), than healthy individuals (Fig. 3C, D). This both reaffirms and extends our findings by ELISpot, demonstrating 1) that rapamycin promotes the formation antigen-specific memory T cells (rather than increasing IFNγ competency), and 2) affects both the CD4⁺ and CD8⁺ T cell compartments.

Memory T cell subpopulations within both the CD4⁺ and CD8⁺ compartments, defined by expression of CD45RA and CCR7, have overlapping and distinct roles in immunity, and in protection from SARS-CoV-2 in particular (27–31). We therefore assessed the frequencies of Spike-specific CD4⁺ and CD8⁺ T cells of a central memory (T_CM; CCR7⁺CD45RA^-), effector memory (T_EM; CCR7^-CD45RA^-), terminally differentiated or terminal effector (T_EMRA; CCR7^- CD45RA⁺), and naïve-like (T_N; CCR7⁺CD45RA⁺) phenotypes (Fig. 3B, fig. S5). Rapamycin use did not appear to favor the formation of a particular memory phenotype (Fig. 3G, H), and KTRs on rapamycin exhibited greater frequencies of Spike-specific T_CM, T_EM and T_EMRA cells than healthy controls in both the CD4⁺ and CD8⁺ compartments (Fig. 3E, F). Rapamycin use was also associated with a significantly elevated frequency of Spike-specific CD8⁺CCR7⁺CD45RA⁺ ‘naïve’ T cells relative to healthy controls (Fig. 3E, F). While vaccination does not increase the frequency of antigen-specific naïve T cells, the CCR7⁺CD45RA⁺ phenotype also includes T stem cell memory (T_SCM), a population which maintains the potential to differentiate into the full spectrum of T cell memory, and maintains long-term protection (32).

KTRs receiving everolimus as a component of their immunosuppression demonstrated a significant boost in the frequencies of Spike-specific CD4⁺ and CD8⁺ naïve-like, and CD4⁺ T_EMRA, cells (Fig 3E, F), but not in the total frequency of Spike-specific CD4⁺ and CD8⁺ T cells (Fig 3C, D).

These data suggest that rapamycin confers an improvement in the quantity of memory T cells that form in response to vaccination in humans. This effect was evident in both the CD4⁺ and CD8⁺ linages, and resulted in quantitative increases in the three major memory T cell compartments with defined roles in antiviral immunity. This trend was also evident at three months post-second vaccine dose in a subgroup of participants who had not received their third COVID-19 vaccine dose at follow-up (fig. S7).

Spike-specific memory T cells associated with rapamycin use are highly functional

Previously, low-dose rapamycin treatment during the T cell contraction phase has been shown to improve the functionality of CD8⁺ memory T cells that arise in response to viral challenge - an effect that is uncoupled from the increase in memory CD8⁺ T cell frequency (17). Conversely, however, higher dose rapamycin, or genetic ablation of mTORC1, impairs the CD8 effector recall response (17, 33). We therefore assessed the functionality of Spike-specific memory T cells in the mTORi-treated KTRs compared to healthy individuals. Production of antiviral effector molecules IL-2, IFNγ, TNF, granzyme B (GZMB) and perforin (PRFN) in response to Spike peptide stimulation was assessed by intracellular cytokine staining (ICS) (Fig. 3I-R).

Functional Spike-specific T cells were enumerated as CD4⁺CD154⁺ or CD8⁺ cells producing one or more cytokines (IL-2, IFNγ, TNF) in response to stimulation (34). KTRs receiving rapamycin exhibited a greater frequency of single, double and triple cytokine-producing CD4⁺ T cells compared to healthy individuals, resulting in a marked increase in total frequency of cytokine-producing CD4⁺ T cells (Fig. 3I-L). This was consistent for all cytokines assessed individually (fig. S8). Although the frequency of AIM⁺ CD8⁺ T cells was greater in rapamycin treated KTRs than healthy individuals, the frequency of CD8⁺ T cells expressing each cytokine, and total cytokine positive cells, were similar between the groups (Fig. 3J, fig. S8B). KTRs in the everolimus group had comparable frequencies of functional Spike-specific CD4⁺ and CD8⁺ T cells to healthy individuals (Fig. 3B-L).

Simultaneous expression of multiple cytokines by memory T cells is associated with increased effector function and viral control (35–40). As a measure of functionality, the proportion of cytokine-producing cells that were polyfunctional, and the mean fluorescence intensity (MFI) of cytokines, was compared between healthy controls and KTRs on rapamycin or everolimus. Both mTORi groups demonstrated a similar proportion of cytokine-producing cells that were polyfunctional to healthy individuals (Figure 3Q, R) and, although there was a trend towards increased IFNγ and TNF MFI in CD4⁺ T cells in the rapamycin group, this did not reach significance (Fig. 3O, P). Both mTORi groups exhibited a trend towards lower TNF MFI in CD8⁺ T cells than healthy controls, reaching significance in the everolimus group (Fig. 3P).

As a measure of cytotoxic function, upregulation of intracellular granzyme B (GZMB) and perforin (PRFN) was measured upon stimulation with Spike peptides, to evaluate broader cytotoxic capacity of T cells in response to antigen encounter (41). KTRs on rapamycin had increased frequencies of CD4⁺ T cells that upregulated PRFN or PRFN⁺GZMB⁺ in resposne to antigen, while no difference in the frequencies of single or double GZMB/PRFN positive CD4⁺ and CD8⁺ T cells was observed for the everolimus group compared with healthy individuals (Fig. 3M, N; fig. S8).

Combinations of cytokine and effector molecules were further explored (fig. S9). IFNγ and TNF were the most commonly co-expressed molecules by CD4⁺ T cells, and granzyme B and perforin were most commonly co-expressed by CD8⁺ T cells (fig. S9). Penta-functional CD4⁺ T cells were detected in 40% (6/15) healthy individuals and 53% (8/15) KTRs on rapamycin (data not shown). Thus, following COVID-19 vaccination, the expanded pool of Spike-specific CD4⁺ and CD8⁺ memory T cells in KTRs on rapamycin is highly functional.

Immunosuppression switch from MMF to rapamycin is safe and feasible but does not improve T cell response to COVID-19 booster vaccination in KTRs

The importance of T cell response in protection against SARS-CoV-2, and the clear benefit that rapamycin use had on the formation of functional T cell memory in our analysis, prompted the trial of immunosuppression modification with rapamycin as an intervention to enhance deficient T cell responses in KTRs to vaccination. An association between MMF use and severe impairment of seroconversion following COVID-19 vaccination in KTRs has been reported (5). We therefore hypothesized that switching MMF to rapamycin in KTRs on SOC therapy would both alleviate suppression of humoral immunity, and boost T cell responses to a third COVID-19 vaccine dose. The combination of CNI, rapamycin and steroid also minimizes the risk of rejection during the period of immunosuppression alteration relative to the combination of MMF, rapamycin and steroid, but does require a concurrent reduction in CNI dose to avoid nephrotoxicity (42). To test this, we initiated the RIVASTIM-Rapamycin trial (ACTRN12621001412820), a multi-center, randomized, open-label, controlled trial. We enrolled KTRs receiving triple therapy with tacrolimus, mycophenolate mofetil (MMF), and prednisolone, who exhibited a sub-optimal response (anti-RBD IgG<100 U/mL) following two vaccine doses (full criteria presented in S10). Participants were randomized to either cease MMF, start rapamycin 2 mg daily and reduce tacrolimus by 50% (rapamycin switch group), or to remain on existing immunosuppression (no switch control group) (see PRISMA diagram, fig. S11). Study protocol, methods and outcomes were pre-specified and were published elsewhere (43).

The trial was conducted from November 2021 – April 2022. Fifty-four patients were screened and all met criteria for randomization. A total of 28 patients were randomized to the rapamycin group and 26 to the control group. After attrition, 25 patients in the intervention group, and 23 patients in the control group, were included for outcome analyses. Patients were predominantly male (33/48, 69%), with mean age 59 ± 9.7 years, and had mostly received mRNA primary vaccination courses (28/48, 58%). Full baseline characteristics are shown in Table 3. In the rapamycin group, mean rapamycin trough concentrations were 6.3 ng/mL at time of vaccination, and 5.5 ng/mL at follow-up. Outcomes assessed included immune response to vaccine, including proportion achieving a protective viral neutralization titer threshold (primary outcome measure (24, 43)), anti-spike antibody titer, T cell response, and drug tolerability.

The proportion of KTRs who exhibited protective neutralization titers was not different between groups (p = 0.85): rapamycin group (10/25, 40%) compared to the control group (9/21, 43%) (Fig. 4A). The unadjusted relative risk (RR) with rapamycin switch was 0.93 (95% CI 0.47 - 1.86, p = 0.84). This was also true of live virus neutralization for both ancestral (A.2.2) and Omicron BA.5 variant strains. No significant difference was observed in anti-RBD seroconversion in the rapamycin group compared to controls (12/25 vs 11/23, p = 0.82) (Fig. 4B), and this persisted after adjustment for baseline anti-RBD level (fig. S12, Table S3).

• Download figure
• Open in new tab

Fig. 4. Immunological results of the RIVASTIM-Rapamycin trial.

KTRs on SOC triple therapy who responded poorly (anti-RBD IgG<100 U/mL) to a primary two-dose vaccination schedule were randomized to the control ‘no switch’ or treatment ‘rapamycin switch’ trial arm four weeks prior to third dose of BNT162b2 COVID-19 vaccine. (A) Serum neutralization of live SARS-CoV-2 (ancestral A.2.2 strain) expressed as log IC_50. (B) Spike-specific T cell responses pre- and post-booster vaccination, as measured by IFNγ ELISpot (SFU/10⁶ cells). There was a significant increase in response in both groups (p < 0.05). (C) Change in Spike-specific T cell response, as measured by change in IFNγ ELISpot (SFU/10⁶ cells). There was no significant difference in median change in T cell responses between ‘no switch’ and ‘rapamycin switch’ groups by quantile regression (p = 0.89). (D) Comparison of post-vaccination anti-RBD IgG titer (U/mL) between ‘no switch’ and ‘rapamycin switch’ cohorts. There was no significant difference in the proportion of patients reaching the pre-defined target threshold of 100 U/mL between groups (χ² = 0.03, p = 0.87). ns, non-significant.

T cell responses as measured by IFNγ ELISpot increased following vaccination in both groups (p < 0.05, Fig 4C). However, the magnitude of change in T cell response did not differ between groups (p = 0.89, Fig 4D).

There was no significant difference in the frequency of any adverse events between the switch and control groups (27% vs 9%, p = 0.10), and there was no significant change in estimated GFR, and no episodes of transplant rejection over the course of the trial (Table S1-S2). Full details of secondary outcomes are available in the Supplementary Materials.

In summary, switching KTRs from MMF to rapamycin prior to a booster dose of COVID-19 vaccine was feasible and well-tolerated, however did not produce a positive effect on T cell response, as was apparent after primary COVID-19 vaccination for KTRs receiving rapamycin as a component of maintenance immunosuppression.

Rapamycin enhances the frequency and function of antigen-specific memory T cells in mice

Immune response to booster vaccination is dominated by memory recall rather than de novo activation of naïve T cells (44). This, coupled with earlier reports that rapamycin influences memory formation by affecting the short-lived effector cell (SLEC) versus memory-precursor effector cell (MPEC) fate decision following activation of naïve T cells (17, 19), prompted us to investigate if rapamycin treatment differentially affects primary versus booster vaccine responses. To do so, we comprehensively evaluated the effect of rapamycin treatment before primary or booster COVID-19 vaccination in Balb/c mice (Fig. 1). Five groups of mice (n=7/group) were vaccinated with two doses of BNT162b2 vaccine (1 _μg/dose, IM), treated with rapamycin before primary dose, or before the third (booster) dose. To compare the effect of rapamycin on naïve versus recall response, mice were boosted with homologous (BNT162b2) or with plasmid DNA vaccine encoding Omicron SARS-CoV-2 variant RBD (pVAX-RBD-Omicron) (Fig. 1 and Fig. 5A).

• Download figure
• Open in new tab

Fig. 5. The effect of rapamycin on the immunogenicity of SARS-CoV-2 original and variant-strain vaccines is influenced by timing of treatment.

(A) BALB/c mice received a primary vaccination schedule of BNT162b2 (Pfizer-BioNTech), followed by a booster dose of homologous (BNT162b2) or 50 ug of Omicron-specific (pVAX-RBD-Foldon[Omicron]) DNA vaccine. Mice were injected daily i.p. with 75μg/kg rapamycin beginning 5 days prior to the first dose (‘pre-primary’), or 5 days prior to booster vaccination (‘pre-booster’), or were not treated with rapamycin (‘no treatment’). Splenocytes were stimulated for 18 hours with peptides derived from the full-length Spike protein, and antigen-specific T cells defined by co-expression of the activation-induced markers (AIM) CD25 and CD69. (B-F) AIM⁺ T cells of (B) naïve (CD44^lowCD62L⁺), (C) central memory (CD44^highCD62L⁺), (D) effector memory (CD44^highCD62L^-), and I double negative (CD44^lowCD62L^-) phenotypes, and total AIM⁺ cells (F), were quantified as a percentage of CD4⁺ and CD8⁺ T cells. (G-H) Mean fluorescence intensity (G) and percentage (H) of stem cell antigen (SCA)-1 expression by AIM^- naïve-like (CD44^lowCD62L⁺) T cells (i) and AIM⁺ naïve-like (CD44^lowCD62L⁺) T cells from no treatment (ii), pre-primary (iii), and pre-booster (iv) rapamycin treated mice. I-N: Functional antigen-specific CD4⁺ and CD8⁺ T cells were assessed by expression of IFNγ and TNF following 16-hour stimulation with peptides derived from the full-length SARS-CoV-2 ancestral-strain Spike protein. (I-J) Percentage of CD4⁺ (I) and CD8⁺ (J) T cells expressing at-least one cytokine in response to stimulation. (K-L) Cytokine profile (outer pie chart) and polyfunctionality (inner pie chart) of Spike-specific CD4⁺ (K) and CD8⁺ (L) T cells. (M-N) Mean fluorescence intensity of cytokine expression by IFNγ⁺ and TNF⁺ T cells in the CD4⁺ (M) and CD8⁺ (N) compartments. (O-P) The effect of rapamycin treatment beginning pre-booster dose on wildtype (O) and Omicron-specific (P) T cell responses to the BNT162b2 and pVAX-RBD-Foldon[Omicron] vaccines, respectively. (Q) Serum titers of anti-RBD IgG following the first (D1), second (D2) and third (D3) vaccine doses. (R) Effect of rapamycin treatment from pre-primary or pre-booster time points on the final (D3) anti-RBD IgG titer. Multiple comparisons by one-way ANOVA with Holms’ correction: ns, non-significant; *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

Supporting the findings in KTRs, administration of rapamycin from baseline resulted in significant improvement in the frequency of functional Spike-specific T cell memory populations following three doses of BNT162b2 vaccine (Fig. 5). Similar to patients, rapamycin treated mice had significantly higher frequencies of Spike-specific CD4⁺ and CD8⁺ T cells by AIM assay (Fig. 5F; fig. S13). This increase was evident for CD4⁺ T_EM and CD44^lowCD62L^- populations, and for CD8⁺ T_EM and T_CM populations (Fig. 5C, D). Increased frequencies of AIM⁺ cells of a naïve-like (CD44^lowCD62L⁺) phenotype were observed in both the CD4⁺ and CD8⁺ T cell compartments, mirroring the increased frequencies of AIM⁺ naïve-like (CD45RA⁺CCR7⁺) CD8⁺ population in KTRs on rapamycin (Fig. 5B; Fig. 3E, F). In contrast to the bulk (AIM^-) naïve-like T cells, this antigen-specific population expressed high levels of SCA-1, suggestive of a T_SCM identity (Fig. 5G, H). Stem-like memory T cells have also been reported to express additional memory markers such as CXCR3; AIM⁺CD44^lowCD62L⁺SCA-1⁺ cells in our study were negative for CXCR3, and thus embody a partial T_SCM phenotype (fig. S14). mTORC1 inhibition has been reported to increase T_SCM formation in vitro (45), and the percentage of AIM⁺CD44^lowCD62L⁺ cells positive for SCA-1, and the MFI of SCA-1 staining, was significantly higher in rapamycin treated mice (Fig. 5G, H).

Rapamycin treated mice exhibited increased frequencies of CD4⁺ and CD8⁺ cytokine-producing T cells (Fig. 5I, J), as well as increased frequencies of multi-functional (TNF⁺IFNγ⁺) CD4⁺ and CD8⁺ T cells (Fig. 5L, K). Similar to KTRs (Fig. 4R), rapamycin favored the formation of IFNγ single-positive T cells (Fig. 5L, K). This resulted in a smaller proportion of cytokine-producing CD8⁺ T cells being multi-functional (TNF⁺IFNγ⁺) despite the increased frequency of these cells (Fig. 5L, K). In support of an additional beneficial effect of mTORC1 inhibition on memory T cell functionality, rapamycin treatment increased the average MFI of both TNF and IFNγ expression by cytokine positive CD4⁺ T cells (Fig. 5M). Of note, while it did not reach significance, a small reduction in the MFI of TNF expression by TNF⁺CD8⁺ T cells was observed that mirrored a similar finding in the KTRs receiving mTORi compared with healthy individuals (Fig. 5N; Fig. 3P).

Considerable variation in T cell response was observed in the rapamycin-treated group, particularly in the CD8 compartment (Fig. 5). Rapamycin concentrations were measured in serum at the time of T cell analysis and were found to inversely correlate with CD8⁺ T cell response by AIM (r = −0.8696, p = 0.0244) and cytokine production (r = −0.9346, p = 0.0063) (fig. S14). This negative association is consistent with reports that high dose rapamycin or genetic ablation of mTORC1 impair memory T cell function at recall (17, 33).

To evaluate the effect of rapamycin treatment on the humoral immune response in mice, we measured anti-RBD IgM and IgG titers. Following the first vaccine dose, mice receiving rapamycin demonstrated significantly higher titers of anti-RBD IgM, and lower titers of anti-RBD IgG (fig. S14), in line with previously reported effects of rapamycin (46, 47). These effects diminished considerably, however, with repeat vaccination, such that no significant differences were observed between groups following the second and third doses, for either antibody isotype (Fig. 5; fig. S14).

T cell responses to SARS-CoV-2 original and variant-strain vaccines are influenced by timing of rapamycin treatment in mice

To understand the lack of effect of rapamycin in our randomized, controlled trial, a comparator group of mice were started on rapamycin after two doses of vaccine (prior to administration of the third dose) to evaluate the potential of rapamycin to improve the response to booster vaccination (Fig. 1). Consistent with the conflicting outcome of our observational and interventional studies in KTRs, beginning rapamycin treatment prior to booster vaccination produced a modest, but non-statistically significant, increase in the frequency of Spike-specific CD4⁺ and CD8⁺ T cells upregulating activation markers and antiviral cytokines in response to stimulation (Fig. 5F, I, J). Significant increases were observed in Spike-specific naïve-like T cells in the both CD4⁺ and CD8⁺ compartments, as well as elevated SCA-1 expression in this population, indicating a partial effect relative to rapamycin treatment from baseline (Fig. 5B, G).

As rapamycin is thought to influence the fate of activated naïve T cells (17), a predominant process of primary rather than recall immunity (48), we next evaluated if the effect of rapamycin is more pronounced for novel antigens. The implication of this, given that the majority of the population has been exposed to SARS-CoV-2 via vaccination or infection, would be that rapamycin treatment could preferentially enhance the response to novel epitopes in variant vaccines. We therefore assessed the response to an Omicron variant RBD plasmid DNA vaccine (pVAX-RBD-Foldon[Omicron]) in mice that had received two doses of BNT162b2 vaccine (Fig. 1). The effect of rapamycin on the naïve T cell response (response to Omicron-specific epitopes in Omicron-boosted mice) and the recall T cell response (response to equivalent ancestral epitopes in homologous boosted mice) were compared. Responses were assessed by IFNγ ELISpot to maximize sensitivity for detection of novel epitope-specific T cells.

Beginning rapamycin treatment prior to boosting with the Omicron RBD vaccine produced a significant 1.58-fold increase in IFNγ-secreting Omicron-specific T cells relative to mice that did not receive rapamycin (Fig. 5P). While this suggests that rapamycin administration at the time of variant booster may benefit the T cell response, a similar (p = 0.4332) 1.43-fold increase in the T cell response to ancestral Spike epitopes was produced by rapamycin treatment of mice peri-booster vaccination with the homologous BNT162b2 vaccine (Fig. 5O). Thus, contrary to the hypothesis that mTORC1 inhibition preferentially enhances T cell responses to novel epitopes, rapamycin had a similar, modest effect on the T cell response to both homologous and variant vaccine boosters.

Collectively, while rapamycin commenced prior to booster vaccination enhanced the T cell response and promoted the formation of T_SCM-like cells, these effects on cellular immunity were significantly greater when rapamycin was commenced prior to initial vaccination.

Study design: observational

Participants were prospectively recruited for the REVAX trial (ACTRN12621000532808), for which a partial cohort was previously published (13). This study was conducted on a single center population of KTRs and their healthy cohabitants commencing in February 2021. Participants were identified by clinical nephrologists and screened. Inclusion criteria were kidney-only transplant over 18 years old with an available cohabitant without kidney disease. Exclusion criteria were patients with past COVID-19 infection, those who had already received a COVID-19 vaccine prior to enrolment, those who could not provide informed consent, or those who did not have a non-immunosuppressed and unvaccinated cohabitant. Demographics including gender, age, cause of kidney disease, and graft details were collected and presented as percentages for ordinal variables, and median ± range for continuous variables.

For AIM and ICS analyses, 15 HC and 15 SOC KTRs were sampled from the complete cohort. This was done using a random number generator (Random.org), which produces true randomness based on atmospheric noise. No experiments were blinded in this study.

Study design: interventional

The RIVASTIM trial was a prospective, multicenter, randomized, controlled trial performed at two academic transplant centers in Australia. Details regarding the trial design and rationale were pre-specified and published previously (43).

SARS-CoV-2 Spike protein production and ELISA

Prefusion SARS-CoV-2 Spike ectodomain (isolate WHU1, residues 1-1208) with HexaPro mutations (kindly provided by Dr Adam Wheatley)(82) and SARS-CoV-2 RBD with C-terminal His-tag (residues 319-541; kindly provided by Prof Florian Krammer) were used in ELISA as we previously described (34, 83). Patient sera were serially diluted and end point titers of SARS-CoV-2 Spike and RBD were calculated and expressed as area under the curve (AUC) using the mean optical density (OD) reading of the first dilution from healthy normal human sera as the baseline cut-off for seropositivity. AUC calculations were performed using Prism GraphPad.

‘Elecsys Anti-SARS-CoV-2 S’ and ‘Elecsys antiSARS-CoV-2’ assays on the Cobas system (Roche), were used to quantify RBD- and nucleocapsid IgG as per manufacturer’s instructions. The quantitation range for detection of anti-RBD IgG in this assay is 0.4-250 U/mL. One U/mL is comparable to 1 BAU/mL.

IFNγ ELISpot

Thawed PBMCs were stimulated for 18h with 4 pools of overlapping peptides spanning the entire length of the spike glycoprotein of the USA-WA1/2020 strain. Individual peptides are 17- or 13-mers, with 10 amino acid overlaps, obtained through BEI Resources, NIAID, NIH (Peptide Array, SARS-Related Coronavirus 2 Spike (S) Glycoprotein, NR-52402).

Briefly, multiscreen-IP HTS plates (Merck Millipore) were coated with anti-human IFN-γ (clone 2G1, Thermo Fisher), and secreted IFN-γ was detected with anti-human IFN-γ biotin (Clone B133.5; ThermoFisher) followed by streptavidin–HRP (BD Biosciences) and AEC substrate (BD Biosciences). Developed spots were counted automatically by use of an ELISpot reader (Cellular Technology Ltd., Bonn, Germany), and the number of spots for unstimulated splenocytes (<50) was subtracted from the number of spots for the peptide pool–stimulated splenocytes to generate the number of specific spot-forming units per 10⁶ cells. Data are presented as mean ± SEM.

SARS-CoV-2 live-virus neutralization assay

Rapid high-content neutralization assay with HEK-ACE2/TMPRSS cells, HAT-24, cells was done as previously described (84, 85). HAT-24 cells were seeded in 384-well plates at 16 × 10³ cells/well in the presence of the live cell nuclear stain Hoechst-33342 dye (NucBlue, Invitrogen) at a concentration of 5% v/v. Two-fold dilutions of patient serum samples were mixed with an equal volume of SARS-CoV-2 virus solution standardized at 2xVE₅₀ and incubated at 37°C for 1 h before adding 40 μL, in duplicate, to the cells. Viral variants used included key variants of concern, Delta (B.1.617.2), Omicron (B.1.1.529) and Omicron sub-variant BA.5, as well as ‘ancestral’ virus (A.2.2) from clade A and presenting no amino acid mutations in Spike (similar to Wuhan ancestral variant). Plates were incubated for 24 h post infection and entire wells were imaged by high-content fluorescence microscopy, cell counts obtained with automated image analysis software, and the percentage of virus neutralization was calculated with the formula: %N = (D-(1-Q)) × 100/D, as previously described (84, 85). Sigmoidal dose-response curves and IC₅₀ values (reciprocal dilution at which 50% neutralization is achieved) were obtained with GraphPad Prism software.

SARS-CoV-2 Spike peptide pools for AIM and ICS

The SARS-CoV-2 peptide pool used for AIM and ICS analyses was kindly provided by Prof Alessandro Sette (La Jolla Institute of Immunology, CA, USA) (86). 15-mer peptides overlapping by 10 amino acids and covering the entire Spike protein sequence were used (total of 253 peptides). All peptides were synthesized and resuspended in DMSO at 1 mg/ml. Thawed PBMCs were rested for 2 h and stimulated with 1 μg/mL of SARS-CoV-2 Spike megapool. PHA 10 μg/mL (Sigma Aldrich) was included as a positive control. An equimolar amount of dimethyl sulfoxide (DMSO, vehicle) was used as a negative control. PBMCs were stimulated for 24 h at 37°C, 5% CO₂, washed and stained with Zombie Green Fixable Live/Dead Stain (L/D, Biolegend) for 20 min, RT, in the dark. Cells were then washed and stained with (CD3 BUV737, CD4 BUV496, CD8 BUV395, CD14 FITC, CD19 FITC, CD45RA BV650, CCR7 (CD197) APC, CD69 PE, CD134 (OX40) PE-Cy7, CD137 (41-BB) BV421) for 20 min, at room temperature in the dark. Fluorescence minus one (FMO) control for antigens: CD45RA, CCR7, CD134, CD69 and CD137 were added to PHA stimulated cells. PBMCs were washed and FACS Fix (0.4%PFA, 20 g/L Glucose, Sodium Azide 0.02% in PBS) was added for 20 min at room temperature, in the dark. Fixed cells were washed, resuspended in FACS wash buffer and data was acquired on BD FACS Symphony. Data analysis was performed using FCS Express ™ (DeNovo Software, Pasadena, CA, USA). All percentages of activated cells were calculated subtracting unspecific DMSO background for each cell phenotype and individual patient.

Intracellular cytokine staining and Spike-specific T follicular helper cell quantification

As described previously (87), PBMCs were thawed and prepared for cell culture as described for the AIM assay. Cells were pre-treated with 0.555 μg/mL of anti-CD40 blocking antibody (HB14, Miltenyi Biotec) for 15 min. Spike peptide pool was then added to a final concentration of 1 μg/mL (making anti-CD40 concentration 0.5 μg/mL for the remainder of the stimulation period). After a 20 h incubation, 2 μM GolgiStop™ (containing monensin, BD, 554724) and 1 μg/mL GolgiPlug™ (containing Brefeldin A, BD, 555029) were added to the cells and incubated for an additional 4 h. Cells were then co-incubated with Fixable Viability stain 780 (BD) and Fc Block (BD) for 20 min, RT, in the dark, washed with FACS buffer solution and stained with surface stain mix (CD3 BUV737, CD4 BUV496, CD8 BUV395, CXCR5 BUV563, CD14 APC-Cy7, CD20 APC-Cy7) for 20 min, RT, in the dark. Cells were washed with PBS and subsequently fixed and permeabilized with Cytofix/Cytoperm™ (BD, 51-2090KZ) for 20 min, RT, in the dark. Cells were then washed with Perm/Wash™ (BD, 51-2091KZ) and stained with ICS stain Mix (CD154 PE, IFNγ PE-Cy7, TNFα APC, PRF1 FITC, IL-2 BV711, GZMB BV421) for 20 min, RT, in the dark. Cells were then washed twice with Perm/Wash™ and once with PBS. Finally, cells were resuspended in PBS and kept at 4°C until data was acquired on BD FACS Symphony. Data analysis was performed using FCS Express ™ (DeNovo Software, Pasadena, CA, USA).

Animals

All mouse work was conducted in accordance with the Australian Code for the Care and use of Animals for Scientific Purposes as defined by the National Health and Medical Research Council of Australia, and studies were approved by the University of Adelaide Animal Ethics Committee. Forty-two female, 6-8-week-old, Balb/c mice were purchased from the University of Monash and randomly allocated to groups (n=7 per group). Balb/c mice were vaccinated intramuscularly with two doses of 1ug of BNT162b2 (Pfizer) vaccine at three-week intervals. Six weeks after the second dose, mice received the third dose of BNT162b2 vaccine (1ug) or 50 ug of pVax-RBD (omicron) DNA vaccine intradermally into the ear pinnae.

Rapamune oral solution (sirolimus, 1 mg/mL) was administered with daily intraperitoneal injection of 0.075mg/kg rapamycin in 100 uL (17). The dosage of rapamycin was adjusted weekly in accordance with average weight of mice in each treatment group. Serum samples were assayed for rapamycin concentrations by liquid chromatography-tandem mass spectrometry using a modification of Koal et al (88). Calibration was performed with matched human serum and whole blood samples, and calibration curves were linear between 0.24 and 4.98 µg/L, with a mean r value of 0.9964 (n=6), quality controls were within 10% of target concentrations and had coefficients of variation of 6.19% at 0.36 µg/L, 3.42% at 1.54 µg/L and 8.89% at 2.64 µg/L.

Mice were culled 3 weeks after the final vaccination and spleen was collected to perform the following assays.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 9.0.0 (San Diego, CA, US) or Stata Statistical Software: Release 14.2 (StataCorp, College Station, TX, US). For human observational studies, no assumptions were made about the distribution of the data sets; non-parametric tests were used in all cases for comparisons. Accordingly, Mann-Whitney and Kruskal-Wallis tests with Dunn’s correction were applied to pair-wise and multiple comparisons, respectively. For analyses of mouse experiments, one-way ANOVA with Holm’s correction was used for multiple comparison between means.

Spearman’s correlation was used to calculate all correlation coefficients. All p-values were corrected using a false discovery rate of 5%.