Mice were bred within the Division of Veterinary Services at Cincinnati Children’s Hospital Medical Center (CCHMC). Strains are available from The Jackson Laboratories [C57BL/6J, 000664; B6.129S7-Asl^tm1Brle/J, 018830; B6.Cg-Tg(Tek-cre)1Ywa/J, 008863; C57BL/6-Tg(H2-Kb-Tcra,-Tcrb)P25Ktk/J, 011005; C57BL/6-Arg1^tm1Pmu/J, 008817 (originally a gift from Peter Murray, St. Jude Children’s Research Hospital); B6.Cg-Tg(Cd4-cre)1Cwi/BfluJ, 022071] and Charles River Laboratories (B6.SJL-Ptprc^aPepc^b/BoyCrCrl, 564).

Complete RPMI 1640 (C-RPMI, 1.14 mM l-arginine, no l-citrulline, 10-040-CV, Corning Cellgro) was prepared by addition of 10% fetal bovine serum (SH30071, GE Healthcare) and 1% penicillin/streptomycin (15140-122, Gibco, Life Technologies). l-arginine-free C-RPMI 1640 (R-free RPMI, 89984, Thermo Scientific) was supplemented with 10% dialyzed fetal bovine serum (35-071-CV, Cellgro, Corning Life Sciences), 1% penicillin/streptomycin (15140-122, Gibco, Life Technologies), and l-lysine (L8662, Sigma-Aldrich) to match the formulation of C-RPMI. R-free C-RPMI was supplemented with l-arginine (A8094, Sigma-Aldrich) or l-citrulline (C7629, Sigma-Aldrich) prepared in sterile water (100 mM stock) at 1 mM final concentrations.

Mycobacterium bovis BCG infection: M. bovis bacillus Calmette–Guérin Pasteur strain was cultured in Middlebrook 7H9 broth (M0178, Sigma-Aldrich) supplemented with 0.05% tween-80 (P4780, Sigma-Aldrich) plus OADC enrichment (R450605, Thermo Fisher Scientific) at 37°C shaking ~50 r.p.m. Bacilli were washed twice with sterile PBS prior to use. For in vitro studies, bacilli were heat-inactivated (HK-BCG) by incubating at 65°C for 30 min and plated on Middlebrook 7H10 agar (262710, Difco) supplemented with OADC enrichment for 3 weeks at 37°C to confirm sterilization. For in vivo infection, anesthetized mice were inoculated with approximately 5 × 10⁶ bacilli by intranasal administration. At 8 weeks postinfection, tissues were harvested and processed for analysis. Infected lung tissue was homogenized in 5 ml sterile PBS and serially diluted on 7H10 agar supplemented with 2.5 mg/l amphotericin B (A9528, Sigma-Aldrich), 200,000 U/l polymyxin B sulfate (P4932, Sigma-Aldrich), 20 mg/l trimethoprim lactate (T0667, Sigma-Aldrich), 50 mg/l carbenicillin (C3416, Sigma-Aldrich), and OADC enrichment. CFUs were quantified following 3 weeks at 37°C. To harvest live mammalian cells, lungs were digested for 1 h at 37°C in DMEM (10-013-CV, Cellgro, Corning Life Sciences) supplemented with 10% bovine calf serum (SH30073.03, Thermo Fisher Scientific), 1% penicillin/streptomycin (15140-122, Gibco, Life Technologies), 0.5 mg/ml deoxyribonuclease I (LS002139, Worthington Biochemical Corporation), and 1 mg/ml collagenase (C7657, Sigma-Aldrich). Lung digests and mLNs were processed into single cell suspensions and stained for flow cytometry. Mtb infection: Mtb Erdman (35801, American Type Culture Collection) was grown in Proskauer–Beck liquid medium containing 0.05% tween-80 to mid-log phase and frozen in 1 ml aliquots at −80°C. Mice were infected with Mtb using an inhalation exposure system (Glas-col) calibrated to deliver 50–100 CFU to the lungs of each mouse, as previously described (26). At day 30 postinfection, mice were sacrificed and lungs were aseptically removed into sterile saline and homogenized. Serial dilutions were plated on 7H11 agar supplemented with OADC. Plates were incubated at 37°C for 3 weeks to enumerate bacterial colonies and calculate bacterial burden.

Mice were injected i.p. with 1 ml sterile thioglycollate (R064710, Thermo Fischer Scientific). Peritoneal exudate cells were collected after 4 days by lavage, followed by red blood cell lysis and plating on 96-well round bottom plates at 1.4 × 10⁵ cells/well. Following adherence, macrophages were stimulated with HK-BCG representing an MOI = 20 to yield consistent T cell stimulation. In some experiments, arginase activity was induced by overnight prestimulation with 10 ng/ml each mouse recombinant IL-4 and IL-10 (14-8041-62 and 14-8101-62, eBioscience). The following day, C-RPMI containing non-adherent cells was aspirated, cells were washed with PBS to remove remaining l-arginine-containing medium, and R-free C-RPMI was added.

Peripheral lymph nodes and spleens were harvested from naïve mice and processed to a single cell suspension. Following red blood cell lysis, lymphocytes were incubated with 2 µM carboxyfluorescein succinimidyl ester (CFSE, Sigma-Aldrich #21888) for 10 min followed by addition of 20% fetal bovine serum v/v in PBS. T cells were plated at 2 × 10⁵ cells/well and stimulated by α-CD3 (BE0002, BioXCell) and α-CD28 antibodies (BE0015.1, BioXCell), p25 peptide (FQDAYNAAGGHNAVF, UBR140905A-1, United Biochemical Research), or HK-BCG-stimulated macrophages in the indicated l-arginine or l-citrulline conditions. In some experiments, cells were supplemented with 2-mercaptoethanol at 55 µM final concentration to promote cell viability. HK-BCG-stimulated wells contained the nitric oxide synthesis inhibitor N-[[3-(aminomethyl)phenyl]methyl]-ethanimidamide, dihydrochloride (1,400 W, 100 µM final concentration, 214358-33-5, Cayman Chemical) to preserve T cell viability, and indicated wells were supplemented with S-(2-boronoethyl)-l-cysteine (BEC, 250 µM final concentration, 63107-40-4, Cayman Chemical) to inhibit arginase activity. T cell proliferation was analyzed by flow cytometry after 72 h in culture. Proliferation index represents the inverse of the mean fluorescent intensity of CFSE within CD4⁺ T cells normalized to the l-arginine/l-citrulline free culture condition.

Single cell suspensions of lymphocytes were obtained as above. Lymphocytes were plated at 2 × 10⁵ cells/well in R-free RPMI with 1 µg/ml α-CD3 supplemented with 1 mM l-arginine, 1 mM l-citrulline, or neither amino acid in the following polarizing conditions (all from eBioscience): T_H1, anti-IL-4 (clone 11B11, 10 µg/ml) and 10 ng/ml mouse recombinant IL-12 p70; T_H2, anti-IL-12 p40 (clone C17.8, 10 µg/ml), anti-IFN-γ (clone R4-6A2, 10 µg/ml), and 5 ng/ml mouse recombinant IL-4; T_H17, anti-IL-4, anti-IFN-γ (clone R4-6A2, 10 µg/ml), 5 ng/ml mouse recombinant IL-6, 2.5 ng/ml recombinant IL-23, and 1.25 ng/ml recombinant TGF-β; regulatory T cell (Treg), anti-IFN-γ, 10 ng/ml TGF-β, and 10 ng/ml recombinant IL-2. Following 5 days of culture, cells were re-stimulated with 10 ng/ml phorbol 12-myristate 13-acetate (P1585, Sigma), 1 µg/ml ionomycin (I0634, Sigma) reconstituted in ethanol, and GolgiPlug (51-2301KZ, BD Biosciences) for 4 h. Cytokine production and transcription factor expression were measured by flow cytometry.

CD4⁺ T cells were isolated from the spleens and peripheral lymph nodes of Asl^flox/+;P25;CD45^.1/.2 and Asl^flox/flox;P25;Tie2-cre;CD45^.2/.2 mice by magnetic sorting with the CD4 isolation kit (130-104-454, Miltenyi Biotec). Following CFSE staining, 1 × 10⁶ cells from each mouse were co-injected at an equal (i.e., 50:50) ratio into C57Bl/6;CD45^.1/.1 mice i.v. One day following transfer, mice were infected with approximately 5 × 10⁶ M. bovis BCG and euthanized at the indicated time points for T cell analysis. mLNs were processed into a single cell suspension and stained for flow cytometry analysis.

In polarization and in vivo experiments, cells were first incubated with a fixable viability dye (65-0865-14, eBioscience) at 1:1,000 dilution in FACS buffer (PBS containing 1% bovine calf serum) for 10 min at room temperature. Samples were blocked using 5% normal mouse serum in FACS buffer for 10 min at 4°C. Cells were stained with antibodies for 30 min at 4°C at 1:200 dilution in FACS buffer (all from eBioscience): CD3 (clone 17A2), CD4 (clone GK1.5), CD45.1 (clone A20), and CD45.2 (clone 104). Following staining, cells were washed and fixed with Foxp3/Transcription Factor Fixation/Permeabilization buffer (00-5521, eBioscience), then stained for intracellular proteins in 1× permeabilization buffer for 30 min at 4°C. The following antibodies were used at 1:100 dilution: IFN-γ (clone XMG1.2), IL-17A (clone eBio17B7), and Foxp3 (clone FJK-16s). Data were acquired using a BD FACSCanto flow cytometer and analyzed with FlowJo software.

Protein RIPA lysates were separated by Tris–HCl buffered 4–15% gradient SDS-PAGE, followed by transfer to Protran membranes. Membranes were analyzed by Ponceau S staining and then blocked in 3% milk in TBS plus 0.05% tween, followed by immunoblot to detect Ass1 (ab175607, Abcam), Asl (PA5-22300, Thermo Scientific), and Grb2 (610112, BD Biosciences).

Supernatants collected from T_H2-polarized cell cultures were incubated overnight with plate-bound α-IL-13 capture antibody (1 µg/ml, clone eBio13A) to allow binding of synthesized IL-13. Plates were then incubated with biotinylated α-IL-13 detection antibody (500 ng/ml, clone eBio1316H) for 30 min, followed by incubation with streptavidin–horseradish peroxidase (E2886, Sigma-Aldrich). Color change of TMB substrate (34021, Thermo Scientific) stopped with 10% phosphoric acid was determined by analyzing the absorbance at 450 nm to quantify IL-13 concentration.

Error bars are the SD or SEM, as described. Data were analyzed for statistical significance by Student’s t-test or two-way ANOVA and p-values are represented as *p < 0.05, **p < 0.01, and ***p < 0.001.

The necessity for l-arginine to promote T cell function is well established (10, 11, 14–16). Most approaches have utilized l-arginine deficiency methods (e.g., l-arginine-free cell culture media, exogenous and/or endogenous arginase activity, etc.) to make this connection. Less focus, however, has been directed toward the impact of l-arginine synthesis (from its precursor, l-citrulline) within T cells, and how this process affects T cell effector functions. As a preliminary step to determine if l-citrulline could be utilized to promote T cell functions, we assessed CD4⁺ T cell proliferation in l-arginine-free RPMI 1640 (R-free RPMI) supplemented with l-arginine, l-citrulline, or neither amino acid. Paralleling previous work (11, 14, 25, 27), T cell proliferation was absent in l-arginine-deficient media compared with that supplemented with l-arginine (Figures 1A–C). When l-citrulline was provided, T cells restored their proliferative ability to the same degree as l-arginine-cultured cells (Figures 1A–C), indicating l-citrulline can rescue T cell proliferation when l-arginine is unavailable. These phenomena are not limited to CD4⁺ T cells, as l-citrulline also promotes CD8⁺ T cell proliferation (Figures S1A–C in Supplementary Material).

We next sought to address the necessity of these amino acids during T cell polarization—a crucial event in driving appropriate effector subsets and subsequent cytokine release essential for cell-mediated immunity. We first analyzed inflammatory cytokine production from CD4⁺ T cells cultured in non-polarizing conditions (T_Ø), T_H1, T_H2, T_H17, or Treg-polarizing conditions in R-free RPMI with l-arginine, l-citrulline, or neither amino acid added. Analogous to proliferation blockade, we observed the production of IFN-γ by T_H1 and IL-17A by T_H17 cells was nearly eliminated when l-arginine was absent (Figures 1D–H). Furthermore, both T_H1 cells and T_H17 cells regained the capacity to synthesize cytokines upon l-citrulline addition (Figures 1D–H). IL-13 production by T_H2-polarized cells also was rescued by l-citrulline in l-arginine-deficient culture (Figure S2A in Supplementary Material). Differentiation of Treg cells, as defined by Foxp3 expression, remained unaltered in the differing amino acid conditions (Figure S3 in Supplementary Material). These data support the need for l-citrulline for T cells to maintain immunological functions when the exogenous l-arginine pool is limited.

l-citrulline is a non-canonical amino acid that is solely metabolized into l-arginine through the sequential activities of Ass1 and Asl, leading us to hypothesize that l-citrulline mediates T cell activity by providing a pool of synthesized l-arginine. To challenge this, lymphocytes from Asl^flox/flox;Tie2-cre mice (Asl^ΔHEM), that are unable to synthesize l-arginine from l-citrulline, were cultured as outlined above in Figure 1. Loss of Asl was verified in Asl^ΔHEM cells by immunoblot, while Ass1 protein expression remained unaffected (Figure S4 in Supplementary Material). Like wild-type T cells, we observed Asl^ΔHEM CD4⁺ T cells divided when provided l-arginine. However, both CD4⁺ (Figures 2A–C) or CD8⁺ T cells (Figures S1D–F in Supplementary Material) cultured in l-citrulline were unable to proliferate. When these CD4⁺ Asl^ΔHEM T cells were polarized and cytokine expression was measured, l-arginine again mediated both IFN-γ and IL-17A production, which was lost in l-arginine-deficient conditions (Figures 2D–H). Unlike wild-type cells, Asl^ΔHEM T cells provided l-citrulline significantly decreased synthesis of IFN-γ and IL-17A. This reduction, however, did not mirror the complete loss of cytokine in the R-free RPMI condition, suggesting an advantage for l-citrulline in cytokine production—but not proliferation—that is independent of l-arginine synthesis. When measured by ELISA, Asl^ΔHEM T_H2 cells were unable to produce IL-13 in l-citrulline media (Figure S2B in Supplementary Material). Altogether, these data reveal functional advantages for l-arginine synthesis from l-citrulline by T cells in a microenvironment where extracellular l-arginine is insufficient.

Thus far, our data show the necessity of l-citrulline when l-arginine is artificially limited, yet l-arginine can also be reduced by the immunoregulatory properties of myeloid Arg1. Arg1 is upregulated in myeloid cells—including macrophages, neutrophils, and myeloid-derived suppressor cells—in response to various conditions, including infections, cancer, and pregnancy (17, 28, 29). Arg1 activity sequesters l-arginine from the microenvironment, creating local competition for the remaining l-arginine pool by surrounding cells. Arg1 activity alone suppresses T cells, and causes T cell hyporesponsiveness (11, 14, 18). This potentially could be detrimental during mycobacterial disease, where anti-mycobacterial T cells might be suppressed by Arg1⁺ myeloid cells, resulting in a blunted host response to infection.

Since our data demonstrate l-citrulline as an alternative source of l-arginine, we next asked if l-citrulline would confer an advantage for T cell function in the face of arginase-mediated suppression during mycobacterial infection. We modeled amino acid competition in vitro to analyze how l-citrulline affects mycobacterium-specific T cell proliferation. Thioglycollate-elicited peritoneal macrophages were primed with heat-killed M. bovis BCG and cocultured with transgenic P25 CD4⁺ T cells, which are specific for antigen 85b of mycobacteria species. This enabled us to model the interaction of mycobacteria-containing macrophages in the lung with mycobacterial antigen-specific T cells. When P25 T cells were cultured with BCG-primed macrophages, they displayed blunted proliferation when cultured in l-arginine when compared with l-citrulline (Figures 3A–C). This was not due to a failure of CD4⁺ T cells to utilize exogenous l-arginine, as they showed a similar proliferative capacity in response to p25 peptide when cultured in l-arginine or l-citrulline, when compared with the polyclonal T cells described in Figure 1 (Figures S5A–C in Supplementary Material). We reasoned that increased Arg1 activity in BCG-primed macrophages accounted for the failure of T cells to proliferate. When we blocked Arg1 activity, either by BEC administration, or using Arg1-deficient macrophages from Arg1^flox/flox;Tie2-cre mice (MΦ^ΔArg1), we observed a restoration in P25 T cell proliferation in the presence of l-arginine (Figures 3D–I). In some experiments when Arg1 was induced in macrophages by IL-4 and IL-10 prestimulation (28), we again observed P25 cells were unable to proliferate in conditions containing l-arginine (Figure S5 in Supplementary Material). Yet, in all conditions when provided l-citrulline, P25 T cells maintained their proliferative ability (Figure 3; Figure S5 in Supplementary Material), demonstrating that T cell l-arginine synthesis from l-citrulline not only supports T cell proliferation, but also escapes Arg1-mediated immunosuppression imposed during mycobacterial and/or cytokine stimulation.

Although these findings provide insight into the potential benefits of l-citrulline over l-arginine during an anti-mycobacterial T cell response, T cells are likely to encounter much smaller concentrations of these amino acids in vivo. The concentrations of free l-arginine and l-citrulline in the lung microenvironment remain to be established; however, in homeostatic serum l-arginine and l-citrulline are maintained at a range of 50–250 µM, prompting us to investigate the effects of l-citrulline and l-arginine at a range of concentrations for T cell proliferation. To do so, P25 T cells were cocultured with BCG-pulsed macrophages as before, but also in titrating concentrations of l-arginine and l-citrulline. Like our previous data (Figure 3), T cell division in 1 mM l-arginine was blunted, presumably due to the arginase activity of the macrophages (Figures 4A–C). Upon addition of l-citrulline, we observed a significant increase in proliferation by T cells supplemented with 1 or 0.1 mM l-citrulline. l-citrulline supplementation also enhanced proliferation in cultures given physiological (0.1 mM), low (0.01 mM), or no l-arginine, although the beneficial effects of 0.1 mM l-citrulline waned once l-arginine availability became low. Taken altogether, these data suggest a physiologically important contribution of l-citrulline for mycobacterium-specific T cell function.

l-arginine restriction during mycobacterial disease poses a challenge for responding T cells to overcome in infected tissues. Previous data have shown that loss of Arg1 leads to improved mycobacterial host defense early, but allows T cell-mediated immunopathology in late stages of infection (21, 30), reminding us that regulating a balance of activation and inhibition is a key property during immune responses. Our data uncover that l-citrulline can replenish l-arginine for T cells, which led us to hypothesize that metabolism of this amino acid would be necessary to sustain T cell-mediated defense against mycobacterial infection in vivo. As a first step to address this, we asked if l-citrulline provided any advantage to T cells early during mycobacterial infection in vivo. CD4⁺ T cells were isolated from P25;Asl^flox/+;CD45^.1/.2 (Asl^WT) and P25;Asl^flox/flox;Tie2-cre;CD45^.2/.2 (Asl^ΔHEM) mice and labeled with CFSE to track proliferation. Cells were then co-transferred into C57Bl/6;CD45^.1/.1 mice at an equal ratio (Figure 5A) in order to distinguish both donor populations from the endogenous lymphocytes (Figure 5B). Mice were then infected with M. bovis BCG and donor cells in the mLN were analyzed prior to infection (day 0) and 4 and 7 days postinfection. Upon analysis of the transferred cells across time, Asl-deficient P25 T cells were reduced in the lymph node compared with Asl-sufficient donor cells (Figure 5C). When CFSE dilution was quantified in these cells, we observed the Asl^WT and Asl^ΔHEM T cells possessed the same capability to divide and similar viability in vivo (Figure 5D; Figure S6 in Supplementary Material), suggesting the difference in frequency was not necessarily due to proliferative capacity or increased death at the time of analysis. Interestingly, Asl-deficient donor cells had reduced cell size and blasting compared with Asl^WT (Figures 5E–G; Figures S6C,E in Supplementary Material), suggesting that l-citrulline metabolism promotes a growth and accumulation advantage to mycobacterium-specific CD4⁺ T cells early during infection.

Since l-arginine synthesis provided an advantage for accumulation of transferred T cells in the draining mLN following infection, we next sought to determine if l-citrulline metabolism was necessary for endogenous anti-mycobacterial T cell accumulation and host defense during infection. To do this, we first developed a model to conditionally delete l-citrulline metabolism in the T cell compartment only. Asl^flox/flox;CD4-cre (Asl^ΔTcell) mice were generated to selectively delete Asl in T cells, which results in a loss of detectable protein by immunoblot in both naïve and activated T cells without affecting other hematopoietic populations (Figure S4 in Supplementary Material). To validate functional deletion of l-arginine synthesis in CD4⁺ T cells from these mice, we analyzed their proliferation in l-arginine or l-citrulline sufficient conditions in vitro. l-arginine supplemented culture enabled CD4⁺ Asl^ΔTcell T cells to proliferate (Figures 6A–C), to a similar extent as wild-type cells in Figure 1. However, those cultured in l-citrulline resulted in loss of proliferation, (Figures 6A–C) mirroring the Asl^ΔHEM T cells (Figure 2) and confirming that l-citrulline metabolism to l-arginine was disrupted in T cells from these mice. Subsequently, Asl^ΔTcell mice and wild-type littermate controls (Asl^WT) were infected with M. bovis BCG to test the necessity of l-citrulline for anti-mycobacterial T cell responses. Surprisingly, upon analysis of mycobacterial burden in the lung, mice lacking T cell l-citrulline metabolism displayed no discernable differences compared with controls (Figure 6D). Similar data were found when infecting these mice with virulent M. tuberculosis (Asl^WT = 6.45 ± 5.13 versus Asl^ΔTcell = 6.21 ± 5.03 log₁₀ CFU, Figure S7 in Supplementary Material). As such, the T cell infiltrate could be analyzed independent of the influence of pathogen burden. Similar to our adoptive transfer data, Asl^ΔTcell mice exhibited reduced frequency of CD4⁺ T cells in the mLNs, and also reduced CD4⁺ T cell accumulation in the infected lung (Figures 6E–H), providing further evidence that l-citrulline metabolism supports T cell accumulation in relevant tissues following mycobacterial infection in vivo.