Mechanistic Target of Rapamycin Inhibitors and Vaccine Response in Kidney Transplant Recipients

Visual Abstract

Abstract

Key Points

• Mechanistic target of rapamycin (mTOR) inhibitor–based immunosuppression was associated with an improved T-cell response to vaccination in kidney transplant recipients.

• Mice treated with an mTOR inhibitor exhibited improved T-cell responses to booster vaccination.

• Switching low and nonresponder kidney transplant recipients to an mTOR inhibitor did not improve T-cell response to a booster vaccination.

Background

Failure to develop protective immunity in response to vaccination is common among kidney transplant recipients, rendering them susceptible to severe infection. Novel strategies are required. Here, we investigated the potential of mechanistic target of rapamycin (mTOR) inhibitors to improve vaccine responses.

Methods

Humoral and cellular responses to primary coronavirus disease 2019 (COVID-19) vaccination (ChAdOx1 or BNT162b2) were assessed for kidney transplant recipients receiving mTOR inhibitor–based (mTOR inhibitor, mycophenolate, prednisolone, n=15) and standard-of-care (tacrolimus, mycophenolate, prednisolone, n=40) immunosuppression, and healthy cohabitants (n=71), in a prospective observational study. Findings were validated and mechanisms explored in mice. Low/nonresponding kidney transplant recipients receiving standard-of-care immunosuppression (N=54) were then randomized 1:1 to switch from mycophenolate to sirolimus or remain on standard of care for 4 weeks before receiving COVID-19 booster vaccination. Augmentation of immunity to COVID-19 was assessed as the primary outcome measure.

Results

A 12-fold greater IFNγ T-cell response to primary vaccination was observed in kidney transplant recipients receiving mTOR inhibitor–based versus standard-of-care immunosuppression (520 versus 43 spot-forming units/10⁶ cells, P < 0.001). A greater frequency of functional memory T cells in the mTOR inhibitor group was observed for both the CD4⁺ (0.20% versus 0.05%, P < 0.001) and CD8⁺ (0.35% versus 0.07%, P = 0.006) compartments by flow cytometry, and kidney transplant recipients receiving mTOR inhibitor–based immunosuppression produced greater frequencies of severe acute respiratory syndrome coronavirus 2–specific CD4⁺ T cells than healthy cohabitants (1.17% versus 0.48%, P = 0.03). In mice, sirolimus treatment enhanced both recall and de novo T-cell responses to homologous and Omicron-specific booster vaccines. Switch from mycophenolate to sirolimus was well tolerated; however, no significant difference was observed in the proportion of kidney transplant recipients in the intervention and control arms that achieved protective virus neutralization (10/25 [40%] versus 9/21 [43%], respectively, P = 0.85) nor in T-cell response to vaccination (P = 0.89).

Conclusions

mTOR inhibition was associated with improved T-cell memory formation in kidney transplant recipients; however, this effect was not reproduced by a short-term mycophenolate to sirolimus switch strategy.

Clinical Trial registry name and registration number:

Australian New Zealand Clinical Trials Registry, ACTRN12621001412820.

Introduction

Kidney transplant recipients are at high risk of severe disease from vaccine-preventable infections, including severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2).^1,2 Immunocompromised individuals were prioritized for vaccination in numerous countries; however, it became clear that standard regimens of coronavirus disease 2019 (COVID-19) vaccines were poorly immunogenic in kidney transplant recipients, and real-world protection from severe disease was limited.^3–12 Vaccine boosters were subsequently found to increase antibody titers in those with poor responses to primary vaccination; however, virus-neutralizing titers remain suboptimal in most patients after a third or fourth dose, and a subgroup of patients remain seronegative.^11–13 A strategy to directly boost vaccine responses in kidney transplant recipients is needed. Such a strategy would likely center on transient modification of immunosuppression and must strike the balance of enhancing vaccine immunogenicity and efficacy without increasing the risk of transplant rejection.

Kidney transplant recipients require lifelong immunosuppression to prevent allograft rejection, and a combination of a calcineurin inhibitor (e.g., tacrolimus), antimetabolite (e.g., mycophenolate mofetil), and steroid (e.g., prednisolone) is the current global standard of care. Alternatively, mechanistic target of rapamycin (mTOR) complex 1 inhibitors, sirolimus (rapamycin) and everolimus, may provide some advantages as an alternative to calcineurin inhibitors in reducing nephrotoxicity¹⁴ or to antimetabolites in reducing risk of viral infection.¹⁵

In preclinical models of vaccination, sirolimus paradoxically promotes the expansion of antigen-specific T-cell memory.^16–18 Treatment of mice or nonhuman primates with sirolimus over the perivaccination period resulted in an increase in the frequency and functionality of the resultant antigen-specific CD8⁺ memory T-cell pool.^16,17 In mice, this effect is mediated by direct inhibition mTOR complex 1 in CD8⁺ T cells and enhances viral control.¹⁷ Consistent with this, mTOR inhibitor use by transplant recipients is associated with reduced viral reactivation and lower incidence of some cancers, although whether these observations result from enhanced T-cell memory is yet to be established.^15,19–21 Importantly, augmentation of the T-cell memory response in mice was selectively observed when T-cell antigens were presented in the context of viral infection and not in the context of transplantation.²² As such, mTOR inhibitors may be uniquely able to enhance antiviral immunity while suppressing the transplant allo-response.

Here, we report superior cellular and humoral immune responses to primary COVID-19 vaccination in kidney transplant recipients receiving mTOR inhibitor–based immunosuppression compared with standard-of-care immunosuppression. Subsequent experiments in mice and a pilot trial of mTOR inhibitor switch in kidney transplant recipients are presented, exploring the mechanism of mTOR inhibitor–augmented vaccine responses and the potential to pursue this strategy clinically.

Methods

Study Approval

Human studies were approved by the Central Adelaide Local Health Network Human Research Ethics Committee, and written informed consent was received from all participants before participation. All mouse experiments were 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.

Study Design: Observational

Participants were prospectively recruited for the REVAX trial (ACTRN12621000532808) from April 1, 2021, until May 30, 2021, for which a partial cohort was previously published.²³ In brief, REVAX recruited outpatient kidney transplant recipients older than 18 years with an available cohabitant without kidney disease. All patients were recruited from the Royal Adelaide Hospital, South Australia. Patients with past COVID-19 infection, those who had already received a COVID-19 vaccine, could not provide informed consent, or did not have a nonimmunosuppressed and unvaccinated cohabitant were excluded.

Patients and cohabitants were vaccinated concurrently, and all peripheral blood samples were collected 3 weeks (median 22 days, interquartile range, 22–25) after the second vaccine dose for analysis of SARS-CoV-2–specific immunity. Humoral immune response to vaccination was measured by total anti-Spike IgG titer, anti–receptor-binding domain (RBD) Ig titer, and live virus neutralization.²⁴ Cellular immunity was assessed by IFNγ ELISpot in response to stimulation with overlapping peptides spanning the SARS-CoV-2 Spike protein, as previously.²⁵ Clinical and demographic information was collated from patient electronic health records.

We compared vaccine responses between kidney transplant recipients and healthy cohabitants and, to account for sources of variation within transplant recipients, stratified by maintenance immunosuppression into patients receiving mTOR inhibitor, mycophenolate, and prednisolone (n=15) versus tacrolimus, mycophenolate, and prednisolone (n=40). Patients who were lost to follow-up or for whom data were missing were not analyzed.

Study Design: Mouse

Forty-two female, 6- to 8-week-old Balb/c mice were purchased from Monash University and housed at the University of Adelaide in a specific pathogen-free facility. Animals were randomly allocated to groups (n=7 per group) and were vaccinated intramuscularly with two doses of 1 µg of BNT162b2 (Pfizer) vaccine at 3-week intervals. Six weeks after the second dose, mice received the third dose of BNT162b2 vaccine (1 µg) or Omicron-specific (pVAX-RBD-Foldon [Omicron]) DNA vaccine (50 µg) intradermally into the ear pinnae. Allocation and treatment were not blinded. Analysis was blinded.

Mice received either control solution (saline) or sirolimus (Rapamune, Wyeth Pharmaceuticals), administered as a daily intraperitoneal injection of 0.075 mg/kg sirolimus in 100 µl.¹⁷ Three groups of mice (n=7 per group) received either daily sirolimus with three doses of BNT162b2 vaccine, vaccine alone, or no treatment. The dosage of sirolimus was adjusted weekly for mouse weight. Serum samples were assayed for sirolimus concentrations by liquid chromatography–tandem mass spectrometry using a modification of Koal et al.²⁶ Mice were culled 3 weeks after the final vaccination, and spleens were collected for flow cytometric analysis. All animals were included for analysis unless euthanized prematurely (n=1).

Study Design: Interventional

The Rapamycin and Inulin for booster VAccine response STIMulation (RIVASTIM) trial was a prospective, multicenter, randomized, controlled trial performed at two academic transplant centers in Australia. Trial rationale, design, and protocol were prespecified and published previously.²⁷ Detailed trial and statistical methods are described in the Supplemental Material.

Statistical Analyses

All statistical analyses were performed using GraphPad Prism 9.0.0 (San Diego, CA), Stata Statistical Software: Release 14.2 (StataCorp, College Station, TX), or R Statistical Software (v4.4.2; R Core Team 2024). For human observational studies, no assumptions were made about the distribution of the variables; nonparametric tests were used in all cases for comparisons. Accordingly, Mann–Whitney U and Kruskal–Wallis tests with Dunn correction were applied to pair-wise and multiple comparisons, respectively. For multivariable linear regression analysis, stepwise regression analysis of biologically plausible variables was used. For analyses of mouse experiments, one-way ANOVA with Holm's correction was used for multiple comparison between means. All P values reported in the text and figures for human observational and mouse experiments were adjusted for multiplicity. Spearman correlation was used to calculate all correlation coefficients using a false discovery rate of 5%.

Results

Study Demographics: Observational Study

Vaccine responses were compared between 15 kidney transplant recipients receiving an mTOR inhibitor in combination with mycophenolate and prednisolone (mTOR inhibitor group); 40 kidney transplant recipients receiving standard-of-care tacrolimus, mycophenolate, and steroids (standard-of-care group); and 71 healthy cohabitant controls (Figure 1A and Supplemental Figure 1). Demographic and clinical details are presented in Table 1. The use of T-cell depleting induction therapy (mTOR inhibitor group, one [7%] versus standard-of-care group, 18 [45%]) and time post-transplant (mTOR inhibitor group, 15 years versus healthy controls, 6 years) was numerically different between groups, while other demographic and transplant criteria were comparable (Table 1).

Figure 1.

Effect of immunosuppression regimen on response of kidney transplant recipients to primary COVID-19 vaccination. (A) Study design. Kidney transplant recipients received two doses of ChAdOx1 or BNT162b2 vaccine, and virus-specific T-cell and antibody responses after 3 weeks were compared between healthy controls (n=71), kidney transplant recipients receiving CNI-based immunosuppression (standard-of-care; n=40), and kidney transplant recipients receiving mTOR inhibitor–based immunosuppression (mTOR inhibitor; n=15). (B) Spike-specific IFNγ-secreting T-cell responses measured as SFUs by ELISpot. (C) Anti-Spike IgG titers in serum. Absorbance was measured for eight serum dilutions and is reported as AUC. Dashed line represents detection threshold (mean AUC+2 SD of negative controls). (D) Anti-RBD Ig titers in serum. The detection range for the assay is 0.4–250 U/ml, as demarcated by the dotted line. Percentage of seropositive individuals (above 0.4 U/ml) is shown for each group. (E) Serum neutralization of live SARS-CoV-2 virus (ancestral strain). Neutralization is reported as the minimum serum dilution required for 50% neutralization. Statistical analysis by Kruskal–Wallis with Dunn correction for multiple comparisons. AUC, area under the curve; CNI, calcineurin inhibitor; COVID-19, coronavirus disease 2019; ELISpot, enzyme-linked immunospot; HC, healthy control; KTR, kidney transplant recipient; MMF, mycophenolate mofetil; mTOR, mechanistic target of rapamycin; mTORi, mTOR inhibitor; Pred, prednisolone; RBD, receptor-binding domain; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; SFU, spot-forming unit.

Table 1.

Baseline characteristics of kidney transplant recipients and healthy cohabitants in the REVAX observational trial

Characteristic	Healthy Cohabitants (n=71)	Standard of Care (n=40)	mTOR Inhibitor (n=15)
Age, yr (µ±SD)	59±12	61±12	64±9
Sex, n (%)	F: 48 (68)	F: 17 (43)	F: 3 (20)
	M: 23 (32)	M: 23 (57)	M: 12 (80)
Vaccine, n (%)	ChAdOx1-S: 36 (51)	ChAdOx1-S: 23 (57)	ChAdOx1-S: 11 (73)
	BNT162b2: 35 (49)	BNT162b2: 17 (43)	BNT162b2: 4 (27)
Cause of kidney failure, n (%)	N/A
GN		17 (43)	9 (60)
Other		9 (22)	2 (13)
Polycystic disease		9 (22)	1 (7)
Renovascular		1 (3)	1 (7)
Diabetes		2 (5)	0 (0)
Unknown		2 (5)	2 (13)
Time post-transplant, yr (µ±SD)	N/A	6±5	15±5
Type of transplant, n (%)	N/A	Deceased donor: 18 (45)	Deceased donor: 10 (67)
		Live unrelated: 13 (33)	Live unrelated: 4 (27)
		Live related: 9 (23)	Live related: 1 (6)
Graft number, n (%)	N/A	First graft: 37 (93)	First graft: 11 (73)
		Second graft: 3 (8)	Second graft: 4 (27)
Blood group, n (%)	N/A	O+ 12 (30)	O+ 5 (33)
		O− 7 (18)	A+ 4 (27)
		A+ 14 (35)	A− 1 (7)
		A− 4 (10)	B+ 3 (20)
		B+ 2 (5)	AB+ 2 (13)
		AB+ 1 (3)
HLA mismatches, n (%)	N/A
0		2 (5)	0 (0)
1		5 (13)	2 (13)
2		5 (13)	1 (7)
3		8 (20)	2 (13)
4		10 (25)	2 (13)
5		7 (18)	5 (33)
6		3 (8)	2 (13)
CMV serostatus, n (%)	N/A	Donor/recipient	Donor/recipient
		D+/R+: 13 (33)	D+/R+: 2 (13)
		D−/R+: 8 (20)	D−/R+: 1 (7)
		D+/R-: 6 (15)	D+/R−: 4 (27)
		D−/R−: 5 (13)	D−/R−: 1 (7)
		Unknown: 8 (20)	Unknown: 8 (53)
Induction therapy, n (%)	N/A	Basiliximab: 19 (48)	Basiliximab: 9 (60)
		ATG: 18 (45)	ATG: 1 (7)
		Eculizumab: 1 (3)	Steroids alone: 2 (13)
		Unknown: 2 (5)	Unknown: 3 (20)
Initial immunosuppression therapy, n (%)	N/A	TAC/MMF/PRED: 39 (97)	CNI/MMF/PRED: 7 (47)
		CyA/MMF/PRED: 1 (3)	CyA/MMF/PRED: 4 (27)
			Belatacept/MMF/PRED: 1 (7)
			Unknown: 3 (20)
Current immunosuppression therapy, n (%)	N/A	TAC/MMF/PRED: 40 (100)	mTOR inhibitor/MMF/PRED: 15 (100)
Comorbidities, n (%)
Hypertension	24 (34)	34 (80)	15 (100)
Diabetes mellitus	10 (14)	16 (40)	4 (27)
Ischemic heart disease	2 (3)	9 (23)	3 (20)
Peripheral vascular disease	0 (0)	2 (5)	0 (0)
Cerebrovascular disease	0 (0)	1 (3)	0 (0)
Chronic lung disease	3 (4)	2 (5)	0 (0)
Dyslipidemia	20 (28)	19 (48)	14 (93)
Cancer (inactive)	4 (6)	5 (13)	2 (14)
Kidney donor	10 (14)	0 (0)	0 (0)
Other	36 (51)	29 (73)	15 (100)
Medications, n (%)
Antihypertensive	21 (30)	34 (80)	15 (100)
Lipid-lowering	20 (28)	19 (48)	14 (93)
Diabetic therapy	8 (11)	13 (33)	4 (27)
Antiplatelet	4 (6)	15 (38)	4 (27)
Anticoagulant	2 (3)	2 (5)	5 (33)
Antiarrhythmic	1 (1)	4 (10)	6 (40)
PPI	19 (27)	26 (65)	10 (67)
Allopurinol	1 (1)	5 (13)	2 (13)
Analgesic	8 (11)	2 (5)	0 (0)
Psychiatric	7 (10)	7 (18)	3 (20)
Other	22 (31)	37 (93)	14 (93)

ATG, anti-thymocyte globulin; CMV, cytomegalovirus; CNI, calcineurin inhibitor; CyA, cyclosporine A; F, female; M, male; MMF, mycophenolate mofetil; mTOR, mechanistic target of rapamycin; N/A, not applicable; PPI, proton pump inhibitor; PRED, prednisolone; TAC, tacrolimus.

Immunosuppression Regimen Influenced Vaccine Responses in Kidney Transplant Recipients

Consistent with previous reports, kidney transplant recipients receiving standard-of-care immunosuppression demonstrated significant impairment in cellular and antibody responses compared with healthy controls (Figure 1, B–E). By contrast, those in the mTOR inhibitor group had a strikingly elevated spike-specific T-cell response, which was 12-fold greater than the standard-of-care group (520 versus 43 spot-forming unit/10⁶ cells, P < 0.001, Figure 1B) and similar to that of healthy controls (520 versus 430 spot-forming unit/10⁶ cells, P = 0.91, Figure 1B). In a stepwise multivariable regression analysis, mTOR inhibitor–based immunosuppression was associated with a 1.56 log₁₀ increase in T-cell response (P < 0.001) among kidney transplant recipients (Supplemental Table 1). Age, HLA mismatches, and comorbid ischemic heart disease were also associated with T-cell response, while comorbid diabetes or previous cancer were not (Supplemental Table 1).

Anti-Spike IgG titers were significantly greater in the mTOR inhibitor group than the standard-of-care group (area under the curve 6.766 versus 0.412, P = 0.01, Figure 1C). The two immunosuppression groups were equivalent in proportion of seroconverted patients with detectable anti-RBD Ig (13 [33%] versus five [33%], P > 0.99, Figure 1D) and in capacity of sera to neutralize live SARS-CoV-2 virus (ancestral strain A.2.2; three [20%] versus five [13%], P > 0.99, Figure 1E). Trough mTOR inhibitor concentrations were above subtherapeutic levels in all participants, and an independent regression analysis did not identify an association with vaccine outcomes for either mTOR inhibitor or tacrolimus concentrations at the time of vaccination (Supplemental Figure 2).

mTOR Inhibition Was Associated with Improved Frequency of Spike-Specific Memory T Cells

To further characterize the elevated ELISpot response in the mTOR inhibitor group, we evaluated the frequency, phenotype, and function of Spike-specific T cells by flow cytometry compared with age-matched standard-of-care and cohabitant controls. Memory T-cell frequency and phenotype were assessed independent of cytokine production by activation-induced marker (AIM) assay, with Spike-specific T cells identified as CD4⁺CD134⁺CD137⁺ or CD8⁺CD69⁺CD137⁺ (Supplemental Figure 3).^28,29 Consistent with the elevated IFNγ ELISpot response in this group, the median frequency of Spike-specific CD4⁺ T cells in the mTOR inhibitor group was 9.0-fold greater than standard-of-care (1.17% versus 0.13%, P < 0.001) and 2.4-fold greater than healthy controls (1.17% versus 0.48%, P = 0.034; Figure 2A). Similar increases were observed for the CD8⁺ T-cell response, with median frequencies of Spike-specific CD8⁺ T cells in the mTOR inhibitor group 5.9-fold higher than standard of care (0.48% versus 0.08%, P = 0.033) and a trend toward greater frequencies than healthy individuals (0.48% versus 0.28%, P = 0.43; Figure 2B). Thus, mTOR inhibitor–based immunosuppression was associated with greatly improved formation of CD4 and CD8 T-cell memory relative to standard-of-care immunosuppression and with greater CD4⁺ T-cell memory responses than healthy individuals.

Figure 2.

Kidney transplant recipients on mTOR inhibitor–based immunosuppression formed higher frequencies of virus-specific CD4⁺ central and effector memory T cells. (A and B) Percentage of AIM-positive (AIM⁺) cells within the CD4⁺ (A) and CD8⁺ (B) T-cell compartments. (C and D) Frequency of each AIM⁺ subpopulation as a percentage of the total CD4⁺ (C) or CD8⁺ (D) T-cell 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⁺. Statistical analysis by Kruskal–Wallis with Dunn correction for multiple comparisons. AIM, activation-induced marker.

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.^30–34 We therefore assessed the frequencies of Spike-specific CD4⁺ and CD8⁺ central memory (T_CM; CCR7⁺CD45RA⁻), effector memory (T_EM; CCR7⁻CD45RA⁻), terminally differentiated (T_EMRA; CCR7⁻CD45RA⁺), and naïve-like/stem-like (T_N/SCM; CCR7⁺CD45RA⁺) T cells. Kidney transplant recipients in the mTOR inhibitor group exhibited higher frequencies of Spike-specific T_CM, T_EM, and T_EMRA cells over standard of care in both the CD4⁺ and CD8⁺ compartments (Figure 2, C and D). Abnormal skewing of memory subsets was observed in the standard-of-care group but not the mTOR inhibitor group (Supplemental Figure 4).

mTOR Inhibition Was Associated with Improved Function of Spike-Specific Memory T Cells

The functionality of Spike-specific memory T cells was assessed by intracellular cytokine staining (ICS) as CD4⁺CD154⁺ or CD8⁺ cells producing one or more cytokines (IL-2, IFNγ, TNF) in response to stimulation (Supplemental Figure 3).³⁵ As seen in Figure 3A, mTOR inhibitor–treated kidney transplant recipients exhibited a greater mean frequency of total cytokine-producing CD4⁺ T cells than standard of care (0.20% versus 0.05% P < 0.001) and similar to healthy individuals (0.20% versus 0.10%, P = 0.27). Similarly, mTOR inhibitor–treated kidney transplant recipients had a higher mean frequency of total cytokine-producing CD8⁺ T cells than standard of care (0.35% versus 0.07%, P = 0.006, Figure 3B) and similar to healthy individuals (0.35% versus 0.54% standard of care P > 0.99, Figure 3B). Findings were similar across single, double, and triple cytokine-producing CD4⁺ and CD8⁺ T cells (Figure 3, C and D).

Figure 3.

Kidney transplant recipients on mTOR inhibitor–based immunosuppression formed highly functional virus-specific T-cell memory. (A and B) Frequency of CD4⁺CD154⁺ (A) and CD8⁺ (B) T cells upregulating at least one cytokine in response to stimulation, as a percentage of total CD4⁺ and CD8⁺, respectively. (C and D) Frequency of Spike-specific CD4⁺CD154⁺ (C) and CD8⁺ (D) T cells single, double, or triple positive for any combination of IL-2, TNF, and IFNγ cytokines, as a percentage of total CD4⁺ and CD8⁺, respectively. (E and F) Percentage of CD4⁺ (E) and CD8⁺ (F) T cells positive for intracellular GZMB (GZMB⁺) and PRFN (PRFN⁺) after stimulation with Spike peptides. (G and H) Cytokine expression profile (outer pie chart) and proportion of cytokine-positive cells that are polyfunctional (inner pie chart) of CD4⁺CD154⁺ (G) and CD8⁺ (H) T cells. Statistical analysis by Kruskal–Wallis with Dunn correction for multiple comparisons. GZMB, granzyme B; PRFN, perforin.

Upregulation of intracellular granzyme B (GZMB) and perforin (PRFN) was measured on stimulation with Spike peptides, to evaluate cytotoxic capacity of T cells in response to antigen encounter.³⁶ mTOR inhibitor–treated kidney transplant recipients had higher frequencies of CD4⁺ T cells that upregulated PRFN and GZMB in response to antigen compared with both healthy individuals (0.39% versus 0.00%, P = 0.030, Figure 3E) and standard of care (0.39% versus 0.00%, P = 0.033, Figure 3E). Similarly, mTOR inhibitor–treated kidney transplant recipients had higher frequencies of PRFN⁺GZMB⁺ CD8⁺ T cells than standard of care (3.12% versus 0.00% standard of care P = 0.040; versus 2.01% healthy individuals P > 0.99; Figure 3F).

The proportion of cytokine-producing cells that were polyfunctional, and the mean fluorescence intensity (MFI) of cytokines, was assessed as measures of T-cell functionality.^37–41 No differences were observed in the proportion of polyfunctional cytokine-producing CD4⁺ or CD8⁺ T cells between groups (Figure 3, G and H), and an increase in the MFI of cytokine staining was not observed in the mTOR inhibitor group (Supplemental Figure 5).

Sirolimus Improved Functional T-Cell Memory to Vaccination without Impairing Antibody Response in Mice

To further our understanding of these observations in kidney transplant recipients, we used a mouse model of COVID-19 vaccination to investigate the effect of sirolimus administration on immunologic responses to COVID-19 vaccination (Figure 4A).

Figure 4.

Low-dose sirolimus improved functional T-cell memory to vaccination without impairing antibody response in mice. (A) Experimental design. BALB/c mice received three doses of BNT162b2 vaccine. Mice were injected daily i.p. with 75 μg/kg rapamune (sirolimus) beginning 5 days before the first dose (RAP) or were not treated with sirolimus (NT). (B) Functional antigen-specific CD4⁺ and CD8⁺ T cells were assessed by expression of IFNγ and TNF after 16-hour stimulation with peptides derived from the full-length SARS-CoV-2 ancestral-strain Spike protein. (C) Cytokine profile (outer pie chart) and polyfunctionality (inner pie chart) of Spike-specific CD4⁺ and CD8⁺ T cells. (D) AIM⁺ T cells quantified as a percentage of CD4⁺ and CD8⁺ T cells. (E) Phenotypic profile of AIM⁺ T cells. Naïve-like/stem-like memory: T_N/SCM, CD44^lowCD62L⁺; central memory: T_CM, CD44^highCD62L⁺; effector memory: T_EM, CD44^highCD62L⁻; double negative: T_DN, CD44^lowCD62L⁻. (F) Serum titers of anti-Spike IgM, anti-Spike IgG, and anti-RBD IgG following the first (D1), second (D2), and third (D3) vaccine doses. Multiple comparisons by one-way ANOVA with Holms correction. i.p., intraperitoneal; NT, no treatment; RAP, receptor-associated protein.

T-cell response to vaccination was measured by ICS and AIM expression (Supplemental Figure 6). Supporting the findings in kidney transplant recipients, administration of sirolimus resulted in a significant improvement in the frequency of functional cytokine-producing CD4⁺ and CD8⁺ T cells (Figure 4B), with equivalent proportions of monofunctional and polyfunctional cells (Figure 4C). Similarly, sirolimus-treated mice had significantly higher frequencies of Spike-specific CD4⁺ and CD8⁺ T cells (Figure 4D). Sirolimus treatment did not favor the formation of a particular subset of memory T cells, and no significant differences were observed in the proportion of each major memory subset (Figure 4E).

To evaluate the effect of sirolimus treatment on the humoral immune response, we measured anti-Spike IgM and IgG titers and anti-RBD IgG titers. After the first vaccine dose, mice receiving sirolimus demonstrated significantly lower titers of anti-RBD and anti-Spike IgG (Figure 4F), in line with previously reported effects of sirolimus.^42,43 This effect diminished considerably, however, with repeat vaccination, such that no significant differences were observed between groups after the second and third doses (Figure 4F).

Sirolimus Treatment before Booster Vaccination Improved Wild-Type and Variant-Specific T-Cell Responses

As the majority of the population has received multiple COVID-19 vaccine doses, we sought to understand whether mTOR inhibition could improve T-cell responses to a booster vaccine dose, as was observed for primary vaccination. Mice were vaccinated with two doses of BNT162b2 vaccine and then treated with sirolimus or no treatment before receiving a booster vaccine dose (Figure 5A, upper panel).

Figure 5.

Sirolimus treatment before booster vaccination improved ancestral and variant-specific T-cell response. (A) Experimental design. BALB/c mice received a primary two-dose vaccination schedule of BNT162b2, followed by a booster dose of homologous (BNT162b2) or Omicron-specific (pVAX-RBD-Foldon [Omicron]) DNA vaccine. Mice were injected daily i.p. with 75 μg/kg Rapamune (sirolimus) beginning 5 days before booster vaccination (RAP) or were not treated with sirolimus (NT). (B and C) Effect of peribooster sirolimus treatment on antigen-specific CD4⁺ and CD8⁺ T-cell responses assessed by cytokine production (B) and AIM assay (C). (D) Effect of peribooster sirolimus treatment on the frequency of T cells recognizing peptides that are exclusive to the BNT162b2 and pVAX-RBD-Foldon (Omicron) vaccines, respectively. Multiple comparisons by one-way ANOVA with Holms correction. DN, double-negative.

Similar to treatment at the time of primary vaccination, peribooster sirolimus treatment increased the frequencies of CD4⁺ and CD8⁺ T cells producing cytokines in response to Spike antigen (Figure 5B), as well as the frequencies of antigen-specific memory cells by AIM (Figure 5C).

To directly compare the effect of sirolimus on naïve (de novo) versus recall responses, mice were boosted with homologous (BNT162b2) or with plasmid DNA vaccine encoding Omicron SARS-CoV-2 variant RBD (pVAX-RBD-Omicron; Figure 5A, lower panel). T-cell responses to wild-type epitopes (found only in the BNT162b2 vaccine) and Omicron epitopes (found only in the pVAX-RBD-Omicron vaccine) were measured by IFNγ ELISpot. Sirolimus increased the T-cell response to the previously encountered wild-type epitopes (1.6-fold) and the T-cell response to the novel Omicron epitopes (1.7-fold) to a similar extent, consistent with mTOR inhibition improving both de novo and recall responses (Figure 5D).

Taken together, sirolimus improved the function and frequency of antigen-specific T cells after vaccination, agnostic to previous antigen exposure, and without impairment of humoral immune responses.

Study Demographics: Randomized Controlled Trial

To test whether immunosuppression modification with an mTOR inhibitor could improve booster vaccine responses in kidney transplant recipients who were hyporesponsive to primary vaccination, we initiated the RIVASTIM-Rapamycin trial (ACTRN12621001412820), a multicenter, randomized, open-label, controlled trial comparing COVID-19 booster vaccine responses between kidney transplant recipients on standard-of-care immunosuppression or switched from standard of care to a tacrolimus/sirolimus/prednisolone regimen for 4 weeks before and 4 weeks after booster vaccination (Figure 6A and Supplemental Figure 7). The trial was conducted from November 2021 to April 2022. A total of 28 patients were randomized to sirolimus and 26 to the control group. After attrition, 25 patients in the sirolimus group and 23 controls were included for outcome analyses (Supplemental Figure 8 and Supplemental Table 2). Patients were predominantly male (33 [69%]), with mean age 59±10 years, and had mostly received mRNA primary vaccination courses (28 [58%]). Full baseline characteristics are presented in Table 2. The mean sirolimus trough concentration was 6.1 ng/ml in the intervention group at the time of vaccination, while the mean tacrolimus trough concentrations were 5.8 ng/ml and 7.3 ng/ml (P = 0.007) for the sirolimus switch and control groups, respectively (Supplemental Table 3).

Figure 6.

Mycophenolate to sirolimus switch did not improve responses to COVID-19 booster vaccination of kidney transplant recipients. (A) Study design. Kidney transplant recipients on standard-of-care triple therapy who were hyporesponsive to a primary two-dose vaccination schedule were randomized to the control “no switch” or treatment “sirolimus switch” arm 4 weeks before a third dose of BNT162b2 COVID-19 vaccine. (B–D) Primary outcome: serum neutralization of live SARS-CoV-2 ancestral A.2.2 (B) and Omicron BA.5 (C) after booster vaccination expressed as log IC_50. (D) Anti-RBD Ig titer (U/ml) after booster vaccination. There was no significant difference in the proportion of patients reaching the predefined target threshold of 100 U/ml between groups (χ²=0.03, P = 0.87). (E and F) Secondary outcome: spike-specific T-cell responses prebooster and postbooster vaccination, as measured by IFNγ ELISpot (SFU/10⁶ cells). There was a significant increase in response in both groups (P < 0.05). (F) 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 “switch” groups by quantile regression (P = 0.89). Ctrl, control.

Table 2.

Baseline characteristics of kidney transplant recipients in the RIVASTIM-Rapamycin randomized, controlled trial

Characteristic	Sirolimus Switch (n=25)	No Switch (n=23)
Baseline anti-RBD IgG, n (%)
Nonresponder (<0.4 units/ml)	9 (36)	7 (30)
Low-responder (≥0.4 units/ml)	16 (64)	16 (70)
Age, yr (µ±SD)	61±8	56±11
Sex, n (%)
Female	8 (32)	7 (30)
Male	17 (68)	16 (70)
BMI, kg/m 2 , at enrollment
Normal (18.5 to <25)	5 (20)	8 (35)
Overweight (25 to <30)	8 (32)	6 (26)
Obese (30+)	12 (48)	9 (39)
Self-reported ethnicity, n (%)
Aboriginal	0 (0)	1 (4)
Asian	3 (12)	2 (9)
Caucasian	19 (76)	19 (83)
Indian subcontinent	1 (4)	1 (4)
North African/Middle Eastern	2 (8)	0 (0)
Time since transplant, yr
0–5	12 (48)	10 (43)
5–10	7 (28)	5 (22)
>10	6 (24)	8 (35)
Initial COVID vaccine dose, n (%)
BNT162b2 (Pfizer)	14 (56)	13 (57)
ChAdOx1-S (AstraZeneca)	9 (36)	9 (39)
mRNA 1273 (Moderna)	1 (4)	0 (0)
Heterologous vaccination	1 (4)	1 (4)
No. of kidney transplants, n (%)
First graft	20 (80)	18 (78)
Second or greater graft	5 (20)	5 (22)
Current transplant type, n (%)
Deceased	18 (72)	16 (70)
Living	7 (28)	7 (30)
Primary kidney disease, n (%)
Diabetes mellitus	3 (12)	3 (13)
Polycystic kidney disease	7 (28)	4 (17)
GN	7 (28)	8 (35)
Hypertension/renovascular disease	1 (4)	1 (4)
Other	4 (16)	4 (17)
Unknown	3 (12)	3 (13)
Baseline eGFR, ml/min per 1.73 m2 (µ±SD)	55±14	59±20
Baseline UACR ratio, mg/mmol (µ±SD)	5±8	14±23
Baseline UPCR ratio, mg/mmol (µ±SD)a	12±0	5±0

BMI, body mass index; COVID, coronavirus disease; RBD, receptor-binding domain; UACR, urinary albumin-to-creatinine ratio; UPCR, urinary protein-to-creatinine ratio.

^aBaseline urinary protein-to-creatinine ratio was calculated with missing values.

Immunosuppression switch from mycophenolate to sirolimus was safe and feasible but did not improve T-cell response to COVID-19 booster vaccination in kidney transplant recipients.

The primary outcome, proportion of kidney transplant recipients who exhibited protective neutralization of live SARS-CoV-2 virus (ancestral A.2.2), was not different between sirolimus and control groups (10/25 [40%] versus 9/21 [43%], respectively, P = 0.85, Figure 6B). The unadjusted relative risk with sirolimus switch was 0.93 (95% confidence interval, 0.47 to 1.86; P = 0.84). No significant difference between groups was observed in the proportion of participants who achieved target threshold for live virus neutralization of the Omicron BA.5 variant (6/25 [24%] versus 7/21 [33%], P = 0.48, Figure 6C), nor anti-RBD Ig titer (12/25 [48%] versus 11/23 [48%], P = 0.82, Figure 6D), which persisted after adjustment for baseline anti-RBD level (Supplemental Figure 9 and Supplemental Table 4).

T-cell memory, as measured by IFNγ ELISpot, increased after vaccination in both groups (P < 0.05, Figure 6E), with no difference in the magnitude of change between groups (P = 0.89, Figure 6F).

There was no significant difference in the frequency of any adverse events between the sirolimus and control groups (7/25 [28%] versus 2/23 [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 (Supplemental Tables 3 and 5).

Discussion

In this study, we observed a strikingly elevated T-cell response to COVID-19 vaccination in kidney transplant recipients receiving mTOR inhibitor/mycophenolate/corticosteroid, which was 12-fold greater than in patients on standard-of-care immunosuppression and equivalent to healthy individuals. A similar increase in the formation of memory T cells was observed in mice treated with sirolimus perivaccination, suggesting that the elevated response in patients was due to a positive effect of mTOR inhibition. In both humans and mice, the antigen-specific memory T cells that formed in the context of mTOR inhibitor treatment were highly functional and phenotypically normal and may therefore be expected to confer protection against severe disease on real-world virus encounter. A positive effect of mTOR inhibition on the formation of T-cell memory to vaccination has previously been reported in preclinical models^17,44–47; however, to the best of our knowledge, this is the first exploration of this effect in humans.^48–51

A number of strategies have been used to try to improve protective immunity to SARS-CoV-2 in kidney transplant recipients, which predominantly focused on enhancing or maintaining humoral immunity, including multiple vaccine boosters,^{4,6,7,11,12,52,53} as well as ring vaccination,^23,54,55 convalescent plasma and Ig,^56–58 prophylactic mAb therapy,^59,60 and more experimental approaches such as correcting dysregulation of the microbiome.^61,62 Although repetitive booster vaccinations have shown benefit,^5,8,11 there are limitations to this strategy,^63–65 and a subgroup of kidney transplant recipients, particularly those receiving higher doses of antimetabolite (mycophenolate mofetil or mycophenolic acid), remained seronegative.^4,66 Several trials have been conducted that withdrew mycophenolate before booster vaccination with the aim of improving the antibody response in nonresponsive kidney transplant recipients^67–70; however, no benefit to this strategy has been demonstrated in a randomized trial.^69,71

Less attention has been given to improving the T-cell response; however, effective T-cell immunity can confer protection against severe disease from viral infection and cross-protection against antibody-escape variants.^2,35,72 Although the relative contribution of T cells and antibody to protection against infection is a matter of debate, a strong T-cell response was associated with protection from severe disease and death in people lacking neutralizing antibodies.^73–75

The RIVASTIM trial enrolled kidney transplant recipients who responded poorly to the first two vaccine doses and randomized participants to switch from mycophenolate to sirolimus before booster vaccination. As a pilot trial, participant numbers were small and inadequately powered to detect a difference in the primary outcome of neutralization. The trial was, however, sufficiently powered to detect a meaningful difference in T-cell ELISpots between the groups based on our observational findings. Despite this, no significant difference in the magnitude of response from prevaccination to postvaccination, nor in final ELISpot counts, was observed.

A possible explanation for the lack of improvement in T-cell response was the immunosuppression modification strategy tested. The observational study investigated kidney transplant recipients on mTOR inhibitor, mycophenolate, and corticosteroid; however, we elected to switch participants of the randomized trial to a tacrolimus, mTOR inhibitor, corticosteroid regimen, to minimize the risk of transplant rejection, and to relieve suppression of the antibody response by mycophenolate. Mutual antagonism between tacrolimus and sirolimus has been reported due to shared intracellular binding to FK506-binding proteins; however, these drugs are synergistic at the concentrations used in vivo.⁷⁶ Tacrolimus has also been reported to promote the formation of short-lived effector T cells in mice, which occurred at the expense of the formation of memory precursor effector cells.⁷⁷ Another possibility is that the lead-in time for immunosuppression switch was inadequate to establish the beneficial effect of mTOR inhibition or relieve the effect of mycophenolate on lymphocytes. Beginning sirolimus treatment of mice (Figure 5) and nonhuman primates¹⁶ 5 days before a booster vaccination was sufficient to improve the memory T-cell response. Although the t_1/2 of mycophenolate is 9–17 hours, we cannot rule out enduring effects of the drug on lymphocytes in the intervention group leading to overimmunosuppression and prevention of potential beneficial effects of mTOR inhibition.

We acknowledge several other limitations in our study. Owing to constraints on participant recruitment during the height of the COVID-19 pandemic, recruitment targets for the RIVASTIM-Rapamycin trial were not met, and as such, conclusions on the efficacy of mycophenolate to sirolimus switch on the primary outcome of serologic neutralization could not be drawn. In addition, time post-transplant and T-cell induction therapy were potential confounding differences between the immunosuppression groups in our observational study.

In conclusion, we present evidence that mTOR inhibition improves the formation of T-cell memory to vaccination in humans and represents a viable target to restore vaccine responses in immunosuppressed groups. mTOR inhibition at the time of primary vaccination was associated with phenotypically and functionally normal antiviral T-cell responses in kidney transplant recipients, which may provide protection from severe disease and cross-protection against variant strains in these patients who lack effective neutralizing antibody responses.^35,73,78 A pilot trial of switch from mycophenolate to sirolimus to enhance booster vaccination responses did not confer this benefit in kidney transplant recipients on a three-drug regimen. However, the approach was feasible and well-tolerated. mTOR inhibitor treatment in kidney transplant recipients perivaccination should be investigated in an alternative immunosuppression combination, or as a monotherapy in other vulnerable groups, as a short-term therapy to restore or boost vaccine-specific T-cell immunity.

Supplementary Material

jasn-36-2213-s001.pdf ^{(1.7MB, pdf)}

jasn-36-2213-s002.pdf ^{(1.3MB, pdf)}

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/JSN/F243.

Funding

B. Grubor-Bauk: The Hospital Research Foundation Group. P. Hissaria: Health Services Charitable Gifts Board (HSCGB; 70-05-52-05-20). A. Sette and D. Weiskopf: National Institute of Allergy and Infectious Diseases (75N93019C00065).

Author Contributions

Conceptualization: Simon C. Barry, Steven J. Chadban, Cheng Sheng Chai, P. Toby Coates, Branka Grubor-Bauk, Pravin Hissaria, Plinio R. Hurtado, Griffith B. Perkins, Julian Singer, Matthew J. Tunbridge.

Data curation: Cheng Sheng Chai, Pablo Garcia-Valtanen, Christopher M. Hope, Makutiro G. Masavuli, Zelalem Addis Mekonnen, Griffith B. Perkins, Bree Shi, Matthew J. Tunbridge, Arthur Eng Lip Yeow, Tracey Ying.

Formal analysis: Anupriya Aggarwal, Anouschka Akerman, Cheng Sheng Chai, Vanessa Milogiannakis, Griffith B. Perkins, Bree Shi, Matthew J. Tunbridge, Tracey Ying.

Funding acquisition: Simon C. Barry, Steven J. Chadban, P. Toby Coates, Branka Grubor-Bauk, Griffith B. Perkins, Matthew J. Tunbridge.

Investigation: Anupriya Aggarwal, Anouschka Akerman, Cheng Sheng Chai, Pablo Garcia-Valtanen, Christopher M. Hope, Plinio R. Hurtado, Julie K. Johnston, Svjetlana Kireta, Makutiro G. Masavuli, Zelalem Addis Mekonnen, Vanessa Milogiannakis, Griffith B. Perkins, Tania Salehi, Benedetta C. Sallustio, Beatrice Z. Sim, Julian Singer, Shane M. Spencer, Matthew J. Tunbridge, Arthur Eng Lip Yeow, Tracey Ying.

Methodology: Steven J. Chadban, Cheng Sheng Chai, P. Toby Coates, Branka Grubor-Bauk, Christopher M. Hope, Griffith B. Perkins, Tania Salehi, Bree Shi, Julian Singer, Matthew J. Tunbridge, Stuart Turville, Arthur Eng Lip Yeow, Tracey Ying.

Project administration: P. Toby Coates, Christopher J. Drogemuller, Griffith B. Perkins, Tania Salehi, Julian Singer, Matthew J. Tunbridge, Tracey Ying.

Resources: Simon C. Barry, Steven J. Chadban, P. Toby Coates, Iain Comerford, Branka Grubor-Bauk, Pravin Hissaria, Benedetta C. Sallustio, Alessandro Sette, Stuart Turville, Daniela Weiskopf.

Software: George Bouras.

Supervision: Simon C. Barry, Steven J. Chadban, P. Toby Coates, Branka Grubor-Bauk, Pravin Hissaria, Christopher M. Hope, Plinio R. Hurtado, Griffith B. Perkins, Stuart Turville.

Visualization: George Bouras, Griffith B. Perkins.

Writing – original draft: Steven J. Chadban, Cheng Sheng Chai, P. Toby Coates, Branka Grubor-Bauk, Griffith B. Perkins, Matthew J. Tunbridge.

Writing – review & editing: Anupriya Aggarwal, Anouschka Akerman, Simon C. Barry, George Bouras, Steven J. Chadban, Cheng Sheng Chai, P. Toby Coates, Iain Comerford, Christopher J. Drogemuller, Pablo Garcia-Valtanen, Branka Grubor-Bauk, Pravin Hissaria, Christopher M. Hope, Plinio R. Hurtado, Julie K. Johnston, Svjetlana Kireta, Makutiro G. Masavuli, Zelalem Addis Mekonnen, Vanessa Milogiannakis, Griffith B. Perkins, Tania Salehi, Benedetta C. Sallustio, Bree Shi, Beatrice Z. Sim, Julian Singer, Shane M. Spencer, Matthew J. Tunbridge, Stuart Turville, Arthur Eng Lip Yeow, Tracey Ying.

Data Sharing Statement

Anonymized data created for the study are or will be available in a persistent repository upon publication. Clinical Trial Data. Figshare. Anonymized participant-level data for the RIVASTIM-Rapamycin trial are available at figshare.com under DOI: 10.25909/28478474.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/JSN/F244.

Supplemental Methods

Supplemental Figure 1. Study flow diagram for the REVAX observational trial.

Supplemental Figure 2. Association of immunosuppression trough concentration with vaccine response.

Supplemental Figure 3. Gating strategy for quantification of antigen-specific T cells by AIM and ICS assays (human).

Supplemental Figure 4. Phenotype of antigen-specific memory T cells.

Supplemental Figure 5. MFI of cytokine production by SARS-CoV-2–specific memory T cells.

Supplemental Figure 6. Gating strategy for quantification of antigen-specific T cells by AIM and ICS assays (mouse).

Supplemental Figure 7. Overview diagram (RIVASTIM-Rapamycin).

Supplemental Figure 8. Consolidated Standards of Reporting Trials diagram (RIVASTIM-Rapamycin).

Supplemental Figure 9. RIVASTIM-Rapamycin outcome of block randomization.

Supplemental Table 1. Stepwise multivariable regression of factors associated with T-cell response to primary vaccination in kidney transplant recipients.

Supplemental Table 2. RIVASTIM-Rapamycin development of COVID-19 after randomization.

Supplemental Table 3. RIVASTIM-Rapamycin safety reporting.

Supplemental Table 4. Influence of existing immunity on antibody response to booster vaccination (RIVASTIM-Rapamycin).

Supplemental Table 5. RIVASTIM-Rapamycin adverse events after immunization.