Long-term rapamycin treatment reduces a type of immune cell and lowers brain inflammation in aging

In order to assess the broad effect of long-term rapamycin exposure on aging of the immune system, we designed a study to compare mice fed a rapamycin-containing diet (samples named “Old Rapa”) vs mice fed with a diet containing only the encapsulating compound EUDRAGIT® (“Old Eud”). The diet composition and rapamycin dose were selected based on the original study by Harrison et al., demonstrating lifespan extension following rapamycin treatment [16]. Mice were placed on the diet at 17–19 months and assessed 5 months later. Three months old mice fed a standard diet were analyzed simultaneously for comparison (Fig 1A).

Rapamycin-containing diet reduced mTORC1 signaling, as indicated by reduced phosphorylation of ribosomal protein S6 (p-S6) in CD11b+ myeloid cells in the blood. The decrease in p-S6 levels was comparable between two different cohorts, despite differences in age at the start of treatment (8 months vs. 17–19 months), underlying that a similar extent of mTOR inhibition can be achieved at this dose and treatment duration, irrespective of age (Fig 1B).

In order to assess the effect of rapamycin on proportions of peripheral immune cells, we first assessed mature myeloid cells by flow cytometry. Previous publications have highlighted that aging is associated with alterations in immune homeostasis, notably, shifts to a higher myeloid cellular output also called “myeloid skewing” [19,20].

At the study endpoint, blood, bone marrow, spleen, and peritoneal lavage were collected and the percentage of Ly6C + /Ly6G– monocyte and Ly6C + /Ly6G+neutrophil populations was assessed (Fig 1C). Monocyte and neutrophil proportions were quantified as percentage of live cells.

In our cohort, we did not observe an effect of rapamycin on monocytes or neutrophils in any of the tissues investigated (Fig 1D) or an increase of the percentage of myeloid cells with age (Fig 1D).

These data collectively suggest that rapamycin does not exert any significant effect on monocyte and neutrophil numbers, as their percentage in blood, bone marrow, peritoneal lavage and spleen in aged mice was not affected.

We next investigated if a rapamycin-containing diet could impact the lymphoid lineage, focusing first on B cells in the bone marrow, lymph nodes, peritoneal lavage, and blood (Fig 2A).

B-cell differentiation in the bone marrow was assessed by measuring cells expressing B220 and the μ-chain of immunoglobulin M (IgM). No significant changes were observed in the proportions of pro-B cells, pre-B cells and immature B cells (Fig 2B). However, although not statistically significant, we noted a trend towards an increase in mature and transitional B cells in aged mice receiving rapamycin, resembling levels observed in young mice (Fig 2B).

Regarding naïve B cells in lymph nodes and spleen, no substantial differences were noticed across the three groups (Fig 2C).

Finally, we investigated B1 and B2 cell populations in the spleen and peritoneal lavage based on IgM, CD11b and CD5 markers. Consistent with previous findings, no significant differences were detected across B cell subpopulations, including subsets of B1 cells such as B1a and B1b, in the spleen and in the peritoneal lavage (Fig 2D).

Altogether, these results indicate that low-dose rapamycin administration has a minimal impact on B cells in aging.

Aging is associated with a decline in naïve T cell numbers and an increase in memory T cells [21]. In previous studies, rapamycin has been shown to increase differentiation of memory CD8 T cells [22,23], however, mTOR-deficiency impairs differentiation of CD4 T cells to Th1, Th2 and Th17 subsets [24]. Therefore, we assessed whether rapamycin treatment could influence the age-associated skewing of naïve to memory T cell populations. To this end, we assessed naïve (CD62L+CD44–), central (CD62L+CD44+) and effector memory (CD62L–CD44+) CD4 and CD8 T cells by measuring the expression of lymph node homing receptor, CD62L, a marker for naïve and central memory T cells, together with CD44, a marker of T cell memory and activation. Helper (CD4+) and cytotoxic (CD8+) T cells were assessed from lymph node, spleen and peritoneal lavage (Fig 3A).

CD4 + T cells isolated from lymph node and spleen of rapamycin-fed mice showed a slight increase in naïve T cells in lymph nodes and spleen, as well as an increase in central memory CD4 + T cells in the spleen (Fig 3B, D).

No differences were observed for CD8 ⁺ T cells in the lymph node, albeit an increase was observed in CD8 ⁺ central memory T cells in the spleen of rapamycin-fed mice compared to both the young and old Eudragit groups (Fig 3E). This corroborates previous literature reporting that mTORC1 inhibition via rapamycin can induce an increase in the generation of CD8 + memory T cells [25].

In the peritoneal lavage, CD4 + T cells showed a switch from naïve to effector T cells (Fig 3F), which appeared specific to the rapamycin treatment. To a certain extent, CD8 + T cells in the peritoneal lavage showed a similar trend, although it did not reach statistical significance (Fig 3G).

Overall, these data show modest changes in T cell populations and differentiation to CD8 + T memory cells following rapamycin treatment, in line with previous reports.

While the effect of rapamycin on T cells expressing the main alpha/beta T cell receptor have been investigated, the effect on the less abundant gamma-delta T cell population, which express a type of T cell receptor that does not require MHC recognition, are not well described. In the context of cancer, rapamycin has been shown to enhance γδ T cell responses [26], however, the effect of long-term rapamycin administration on γδ T cell proportion remains unknown. For this reason, we next assessed whether rapamycin diet could have an effect on γδ T cell subsets.

Peritoneal γδ T cells are subdivided in IFN-γ- and IL-17-producing cells, distinguished by the expression of surface marker CD27 [27]. CD3 + , CD4–, CD8–, γδ TCR + , CD27+ and CD27– cells were assessed in the lymph node, spleen and peritoneal lavage (Fig 4A-D).

In the peritoneal lavage, CD27 + γδ T cells showed a decrease, while CD27– γδ T cells showed an increase with age (Fig 4D). In contrast, while rapamycin did not affect CD27 + γδ T cells, it significantly affected the IL-17 producing, CD27– counterparts, rescuing their levels to those observed in young mice (Fig 4D). Similar trends were observed in the lymph node and spleen, although differences did not reach statistical significance (Fig 4B-C).

Taken together, these data indicate that rapamycin prevents the increase in IL-17-secreting γδ T cells that accompanies aging.

Our data indicated that long-term rapamycin did not significantly affect the proportion of peripheral immune cells, except for a clear effect on a subset of IL-17-producing γδ T cells.

Given these differences, we investigated whether systemic levels of inflammatory cytokines were also altered (Schematic in Fig 5A). To this end, serum cytokine levels were assessed, however, we did not observe any differences in serum concentration of IL-17A together with IL-6, IL-10, IL-27 (Fig 5B).

Nonetheless, since IL-17 levels can modulate neuroinflammation [28], we next investigated if the rapamycin diet affected the phenotype of central nervous system (CNS) tissue resident immune cells, microglia. To this end, we assessed the expression of pro-inflammatory cytokines and reactive microglial markers using quantitative reverse transcription PCR (RT-qPCR) in microglia cells. Although aging is known to induce upregulation of inflammatory cytokines and receptors in microglia, our data indicate that rapamycin did not counteract age-associated neuroinflammation (Fig 5C-D).

Although no changes at steady state were observed, we reasoned that systemic and brain inflammation might be affected in response to a strong inflammatory stimulus, such as intraperitoneal exposure to lipopolysaccharide (LPS).

Thus, a separate cohort of mice were fed with either Eudragit or rapamycin diet from 6 months of age for 5 months, then challenged with 2 mg/kg of LPS intraperitoneally and serum cytokine levels and microglia activation were analyzed 4 hours later.

Serum cytokines IFN-β, IL-1β, TNF-α, and IFN-γ were not different between the two diets (Fig 5E). However, the inflammatory cytokines IL-12p40, IL-27 and IL-17A were significantly reduced in rapamycin-fed mice upon LPS stimulation, along with the anti-inflammatory cytokine IL-10 (Fig 5E).

Examining the microglial response, we found that rapamycin administration moderately reduced inflammatory markers Tnf-α, Il6 and Tlr2 gene expression while others, such as Il-12a and Il-12b remained unaffected (Fig 5F).

Taken together, these results suggest that rapamycin diet can in part reduce systemic and brain inflammation in response to a strong systemic inflammatory stimulus.

C57BL6J (RRID:IMSR_JAX:000664) female and male mice of different ages, as specified in figure legends, were purchased from Charles River and maintained on a chow diet containing rapamycin or Eudragit as a control for 5 months. Mice were group-housed under standard conditions at Barts Cancer Institute, Queen Mary University of London, UK and DZNE Bonn, Germany. All experiments were conducted in accordance with relevant institutional and national guidelines, ethically reviewed and approved under UK Home Office Project Licence PPL 70/7411 and German authorization TVA 81–02.04.2018.A215.

Group sizes were determined using the Resource Equation Method and mice were randomly assigned to experimental groups. The rapamycin diet was not associated with observable adverse effects and was not a burden to the animals. Mice receiving i.p. injection of LPS reached a mild and, in some cases, moderate burden, nonetheless they were killed before reaching humane endpoints, due to the short duration of the treatment.

Animals were monitored daily for signs of distress (e.g., hunched posture, reduced mobility, social isolation), which were defined in advance as criteria for humane intervention; however, none of these signs were observed in any experimental animal. All procedures were performed under routine welfare monitoring to minimize pain, distress and suffering. At the end of experiments, mice were either euthanized by CO₂ inhalation, followed by cervical dislocation or anesthetized with isoflurane to collect blood via cardiac puncture. Briefly, mice were placed in an induction chamber with 3% isoflurane (600 ml/min nitrous oxide and 700 ml/min oxygen) for 5–10 minutes and maintained under anesthesia using 2% isoflurane delivered via a nose cone. Adequate anesthesia was confirmed by the absence of withdrawal reflexes prior to sample collection. Cardiac puncture was performed under deep anesthesia, followed by cervical dislocation. No analgesics were administered during the study, as none of the animals required them.

Food-encapsulated rapamycin was prepared following the protocol by Harrison et al. [16]. Rapamycin was obtained from Rapamycin Holdings (now Entomora Biosciences) in San Antonio, Texas, USA. Rapamycin was microencapsulated with an enteric coating material, Eudragit S100 (Röhm Pharma, Germany), which protects rapamycin through the food preparation process. Since Eudragit S100 is water soluble only at non-acidic pH, rapamycin is only released in the small intestine and protected in the stomach. Eudragit-microencapsulated rapamycin and Eudragit as control were incorporated into separate LabDiets on a 5058 base by Purina Labs. The amount of rapamycin used corresponded to 152 ppm, resulting in a 14 ppm concentration of active rapamycin in the diet (2.24 mg of rapamycin per kg body weight).

Female 6-month-old mice were fed a rapamycin- or Eudragit-diet for 5 months, then injected with 2 mg/kg of LPS of E coli 0111:B4 (Sigma, Cat. No. 297-473-0) in 200 μL PBS or PBS only as a control. Intraperitoneal injections were performed using a 27G needle and 1 mL syringe. Mice were sacrificed after 4 hours and organs were collected as described below.

Samples were processed under partial blinding conditions. Following collection, samples were assigned new identifiers and processed by multiple researchers, minimizing awareness of sample identity during downstream handling, although one researcher contributed to both sample coding and part of the processing.

Blood was obtained by cardiac puncture using a 27G needle and 1 mL syringe coated in 0.5 M Ethylenediaminetetraacetic acid (EDTA). It was then transferred into a 1.5 mL EDTA-coated tube and placed on ice before being centrifuged 3 times at increasing speed, removing pellets each time. Centrifugation speeds were 200 x g for 3 min; 2,300 x g for 5 min and 16,100 x g for 3 min. Plasma was then aliquoted and snap frozen before being stored at −80°C until cytokine assessment was carried out.

Spleens were harvested, placed in 10 mL of PBS and kept on ice until further processing. Afterwards, they were placed on a 70 μm cell strainer from Fisher Scientific (Cat. No. 11597522). They were gently mashed through the strainer using a 5 mL syringe plunger. 10 mL of PBS was added to the cell suspension before centrifugation at 300 x g for 10 minutes at 4°C.

Cells were obtained from the bone marrow by flushing the femurs and tibia with Dulbecco’s Modified Eagle Medium (DMEM, Sigma, D5768) supplemented with 10% Fetal Bovine Serum (FBS, Life Technology, Cat. No. 10500−064) and 1% penicillin/streptomycin (Sigma, Cat. No. P4333). 1 mL of complete media was flushed through the bone marrow using a 27G needle, several times. Cells were then strained through a 70 μm cell strainer before 20 mL of PBS was added to the cell suspension and cells were spun at 300 x g for 10 minutes 4°C.

The brain was dissected following perfusion with 35 mL of ice-cold Hank’s balanced salt solution (HBSS) for 5 minutes and subsequently digested using a neuronal dissociation kit (Miltenyi Biotec, Cat. No. 130-092-628). Next, cells were washed twice in HBSS and myelin was removed by negative selection using myelin removal beads (Miltenyi Biotec, 130-094-060) in MACS LS columns (Miltenyi Biotec, Cat. No. 140-096-433). Briefly, 1.8 mL of fluorescence-activated cell sorting (FACS) buffer [phosphate buffered saline (PBS) + 0.5% bovine serum albumin (BSA) + 2mM EDTA] and 200 µL of myelin removal beads were added to each brain cell suspension and incubated at 4°C in the dark for 15 minutes. Following this incubation, 18 mL of FACS buffer was added to wash the cells, then centrifuged for 10 minutes at 300 x g. Using three LS columns per brain, 1 of 3 mL of beads/cell suspension were loaded onto each column placed in a QuadroMACS separator magnet (Miltenyi Biotec, Cat No. 130-042-302). This step was required to separate bead-bound myelin from the cell suspension. Each LS column was washed twice with FACS buffer and each time the flow-through was collected in order to collect brain cells.

Cells from blood, spleen and bone marrow were first subjected to red blood cell lysis using 10X red blood cell lysis buffer (eBioscience, Cat. No. 00-4300-54), diluted 1:10 in distilled water. Cell pellets were resuspended in 10 mL of 1X lysis buffer and placed in the dark for 10 minutes at room temperature. Following this incubation, cells were washed twice with FACSbuffer [phosphate buffered saline (PBS) + 0.5% bovine serum albumin (BSA) + 2mM EDTA] and used for subsequent antibody staining.

Cells were stained in 96-well V-bottom plates from Thermo Fisher (Cat.No. 2605), 2 x 10^6 cells in 100 μL were added to each well and the plate was centrifuged at 300 x g for 5 minutes at 4°C. Cells were then resuspended in 30 μL of CD16/CD32 FcR block (eBioscience, Cat. No. 14-0161-86) for 15 minutes at 4°C. Following FcR block, a 2X antibody mix was prepared using FACS buffer and 30 μL was added to each well. All relevant information regarding the panel of antibodies used (manufacturer, clone and dilutions) is summarized in Table 1. After a 30 min incubation at 4°C in the dark, cells were washed twice with 150 μL of FACS buffer before being resuspended in 150 μL of FACS buffer, which was transferred to 1.2 mL FACS tubes (Star lab, Cat. No. I1412-7400). Data acquisition was carried out with a 4-laser LSRFortessa or a 5-laser FACSymphony (BD Biosciences) on a BD FACSymphony A5 Cell Analyzer (RRID:SCR_022538), using BD FACSDiva Software (RRID:SCR_001456). Subsequent data analysis was done on FlowJo (RRID:SCR_008520).

Following brain homogenization (see Organ harvesting: Brain), the single cell suspension was resuspended in 500 μL of FACS buffer and 40 μL of CD11b microbeads (Miltenyi Biotec, Cat. No. 130-093-636) and incubated at 4°C in the dark for 15 minutes. Following this incubation, cells were washed with 3 mL of FACS buffer and centrifuged for 10 min at 300 x g at 4°C. Using one MS column (Miltenyi Biotec, Cat. No. 130-042-201) per sample, 500 μL of bead/cells were loaded onto each column placed in an OctoMACS separator magnet (Miltenyi Biotec, Cat. No. 130-042-108). Each MS column was washed twice with FACS buffer and then 1 mL of FACS buffer was used to flush CD11b-positive cells using the plunger provided. Purity of microglia was assessed by flow cytometry analysis by staining an aliquot of cells with the microglia-specific markers CD11b and CD45 and reached 92–95%.

Microglia (0.5–1 x 10^6) were lysed in 350 μL RLT lysis buffer (Qiagen, Cat. No. 74004) and stored at −80°C. RNA was extracted with the RNeasy micro kit from Qiagen (Cat. No. 74004), according to the manufacturer guidelines.

Following RNA extraction, RNA was reverse transcribed into cDNA using the High-Capacity cDNA Reverse Transcription kit from Applied Biosystems (Cat. No. 436–8814). Briefly, a master mix was prepared that contained (per reaction): 1X RT buffer, 1X random hexamers, 4 mM dNTP, 1 μL of reverse transcriptase (50 U/ml) and 2.2 μL of nuclease-free water. This master mix was added to the extracted RNA, contained in 12 μL, for a final volume per reaction of 20 μL. All samples were then incubated in a thermocycler using the following program; 25°C for 10 minutes, 37°C for 120 minutes, 85°C for 5 minutes and held at 4°C. Following cDNA synthesis, 130 μL of nuclease-free water was added to each sample to dilute the enzymes used in the cDNA synthesis. A master mix was then prepared for the qPCR for each gene of interest, which included per reaction: 10 μL of Biorad iTaq master mix (Biorad, Cat. No. 436–814) and 1 μL of each FAM labelled gene of interest. This master mix was added to a qPCR plate and 9 μL of diluted cDNA was added to each corresponding well. The Step One Plus Real Time PCR machine from Applied Biosystems was used to carry out the qPCR. The PCR program included an initial denaturation cycle that lasted for 10 minutes at 95°C, followed by 40 cycles of amplification that included 15 seconds at 95°C and one minute at 60°C; the last part of the program included one more cycle at 25°C for 15 seconds. All primers used are summarized in Table 2.

Cytokine levels in serum or in microglia protein lysates were determined using the Legendplex mouse inflammation 13-plex panel, (Biolegend, Cat. No. 740446); a bead-based assay that allows simultaneous measurement of analytes based on cell size using flow cytometry. Specifically, the cytokines assessed were IFN-β, IFN-γ, IL-1β, IL-6, IL-10, IL-12B, IL-17A, IL-27 and TNF-α. The assay was carried out in a 96-well plate following manufacturers’ instructions.

ELISA kits were obtained from BD Biosciences: IL-12p40 (Cat. No. 555165), IL-6 (Cat. No. 555240) and TNFα (Cat. No. 558534). ELISA was performed according to the manufacturer’s guidelines; however, half-surface area 96-well plates were obtained from Corning Costar (Cat. No. 3690) and used for the assay. The wells in these plates are half the volume of standard 96-well plates, allowing the use of a reduced number of samples and reagents. This was particularly important for plasma cytokine levels, as the amount of plasma obtained from each mouse was potentially a limiting factor. Briefly, capture antibody was diluted in PBS and added to each well of a 96-well plate. The plate was then incubated overnight at 4°C, prior to being washed three times with wash buffer (PBS with 0.05% Tween-20). The plate was then incubated with assay diluent (PBS with 10% FBS) for 1 hour, in order to block non-specific binding and washed three times with wash buffer. The higher concentration of standards (1000 pg/mL) were made using a stock solution of each cytokine and then a serial dilution was performed on a separate plate to obtain the remaining standards (500, 250, 125, 62.5, 31.3 and 15.6 pg/mL). Samples and standards were then added to the ELISA plate and incubated for 2 hours at room temperature. Samples that required dilution were diluted in assay diluent. Following incubation, the plate was washed 5 times with wash buffer and then the working detector solution (detection antibody and HRP) was added to the plate and it was incubated at room temperature for 1 hour. The plate was then washed 7 times before the substrate solution [Tetramethylbenzidine (TMB) and hydrogen peroxide from the TMB Substrate Reagent Set (BD PHarmingen, Cat. No.555214)] were added and incubated in the dark for a maximum of 30 minutes. Following the appropriate color change, 25 ml of 1M H₃PO₄ was added to stop the reaction. Plates were analyzed on a plate reader to read the optical density at 450 nm.

Statistical analysis was conducted using GraphPad Prism (RRID:SCR_002798), one-way ANOVA and Welch’s t-test were employed as required and as reported in the respective figure legends. A p value < 0.05 was considered statistically significant. Only P values below 0.1 are reported in the figures.