Cyclophosphamide enhances immunity by modulating the balance of dendritic cell subsets in lymphoid organs

Abstract

Cyclophosphamide (CTX), a commonly used chemotherapeutic agent can enhance immune responses. The ability of CTX to promote the proliferation of effector T cells and abrogate the function of regulatory T cells (Tregs) has been described. In this study, we examined the effects of CTX treatment on dendritic cell (DC) subsets and the subsequent outcome on the effector and suppressive arms of adaptive immunity. In secondary lymphoid tissues, tissue-derived migratory DCs (migratory DCs), lymphoid tissue-resident DCs (resident DCs), and plasmacytoid DCs (pDCs) are well described. CTX has profound and selective cytotoxic effects on CD8(+) resident DCs, but not skin-derived migratory DCs or pDCs in lymph nodes (LNs) and spleen, causing an imbalance among these DC subsets. CTX treatment increases the potency of DCs in antigen presentation and cytokine secretion, and partially inhibits the suppressor activity of Tregs. Adoptive transfer of CD8(+) DCs can reconstitute this population in regional draining LNs and abrogate the immune-enhancing effects of CTX in vivo. These findings demonstrate that CTX may improve immune responses by preferentially depleting CD8(+) lymphoid-resident DCs, which leads to diminished Treg suppression and enhanced effector T-cell function in vivo.

Figure 1

Effect of CTX on the absolute and relative numbers of different DC subsets in lymph nodes. Mice were treated with CTX (150 mg/kg) by intraperitoneal injection and killed 2, 4, 7, and 16 days after drug administration. Subcutaneous LNs and spleen were harvested, digested, and assessed for the number and proportion of skin-derived migratory DCs (CD11c^int, MHC II^high), resident DCs (CD11c^high, MHC^int), or pDCs (CD11c^int, MHC^low, B220⁺, Gr1⁺) at the time points indicated and compared with untreated naive controls. (A) DC subsets in subcutaneous LNs and spleen of naive mice stained for CD4, CD8, and (B) CD86, CD11b, and CD205. Representative plots of naive LN DCs and LN DCs 4 days after CTX treatment (C). The proportions of skin-derived DCs, resident DCs, and pDCs in the LNs as measured by flow cytometry at various time points after the administration of CTX (D). Absolute cell number of different DC subsets was also counted using the total cell count of LN cells. Fold decrease in absolute number of LN DC subsets 4 days after CTX treatment (E). CD8 indicates CD8⁺ resident DCs; CD4, CD4⁺ resident DCs; DN, CD8⁻CD4⁻ resident DCs; and pDC, plasmacytoid DCs. Data are representative of 4 independent experiments. *Statistically significant.

Figure 2

DCs from CTX-treated mice are more potent at antigen presentation. Whole CD11c⁺ DCs from CTX-treated or naive mice were sorted from LNs with anti-CD11c magnetic-activated cell sorting beads. (A) Graded doses of DCs were cocultured with allogeneic T cells from Balb/c mice or were pulsed with OVA peptide and cocultured with antigen-specific CD8⁺ T cells from OT-I mice. During the incubation, MTS/PMS solution was added to the culture to measure the proliferation of T cells. Absorbance at 490 nm was recorded using an enzyme-linked immunosorbent assay plate reader. Representative data from 3 independent experiments are shown. (B) DCs from CTX-treated or naive mice were sorted as described for panel A and cultured in vitro with granulocyte-macrophage colony-stimulating factor for 20 hours with or without stimulation (LPS + IFN-γ). Brefeldin A was present during the last 12 hours of the incubation. (C) DCs were stained with anti-CD11c, MHC class II, B220, Gr-1, and CD8 to identify different DC subsets as defined in Figure 1A and examined for production of IL-12 p40/70. Baseline levels and augmentation of IL-12 p40/70 with the addition of LPS and IFN-γ are depicted. Data are representative of 3 independent experiments. *Statistically significant.

Figure 3

CD8⁺ DCs inhibit the allostimulatory capacity of DCs from CTX-treated mice in vitro. CD11c⁺ DCs from CTX-treated or naive mice were magnetically sorted from LNs. CD8⁺ and CD8⁻ DCs of naive mice were also magnetically or FACS sorted from the spleen. DCs (naive DC indicates DCs from naive mice; CTX DCs, DCs from CTX-treated mice; CD8⁺, CD8⁺ DCs; and CD8⁻, CD8⁻ DCs) were cocultured with the combinations indicated with CFSE-labeled CD3⁺ T cells from allogeneic Balb/c mice for 4 days. (A) CD4⁺ and CD8⁺ T-cell proliferation was examined by measurement of CFSE dilution and the T-cell activation marker, CD44. To assess IFN-γ production, T cells were stimulated with phorbol myristate acetate and ionomycin for 6 hours; brefeldin A was present during the last 5 hours, and then stained for intracellular IFN-γ. Results are the mean ± SEM of triplicate wells. Data are representative of 3 experiments. (B) The proportion, proliferation, and IFN-γ secretion of CD4⁺ Foxp3⁺Tregs were also examined. Results are the mean ± SEM of triplicate wells. Data are representative of 3 experiments. *Statistically significant.

Figure 4

Tregs expanded with DCs from CTX-treated mice have less suppressor activity. Isolated CD4⁺ CD25⁺ Tregs from Balb/c mice were cocultured with different DC subsets (DCs from naive mice, DCs from CTX-treated mice, CD8⁺ DCs, and DCs from CTX-treated mice + CD8⁺ DCs) from B6 mice (primary MLR). After 7 days, Tregs expanded with different DC subsets were mixed with 10⁵ CFSE-labeled CD25⁻ CD3⁺ T cells in various ratios (2:1-32:1) and stimulated with irradiated mature DCs activated with LPS (secondary MLR). At 5 days after secondary MLR culture, T-cell proliferation was examined by measurement of CFSE dilution. Results are the mean ± SEM of triplicate wells. Data are representative of 2 independent experiments. *Statistically significant.

Figure 5

Reconstitution of CD8⁺ DCs in LNs inhibits antigen-specific T-cell responses. Mice were treated with CTX on day −4, followed by the adoptive transfer of 10⁵ CD8⁺ pmel-1 cells from Thy1.1 pmel-1 mice by intravenous injection on day −1. On day 0, mice were immunized with hgp100 peptide–pulsed DCs injected intradermally in the left ear. At 5 days after immunization, draining retroauricular LNs were harvested, and stained for CD8 and Thy1.1 to detect antigen-specific pmel-1 cells (Thy1.1⁺). To assess IFN-γ production, cells were cocultured with mgp100-pulsed EL-4 cells and then stained for intracellular IFN-γ. All mice received adoptively transferred pmel-1 CD8 T cells. As negative control for immunization, mice were not immunized and treated (▴) or not treated (■) with CTX. As negative control for CTX, immunized mice did (▾) or did not (♦) receive CTX treatment. A group of mice received 10⁶ CD8⁺ DCs in the left ear (●) or received 10⁶ CD8⁻ DCs in the left ear (□) on day −1. (A) Experiment timeline. (B) Representative plots. (C) The recovery of antigen-specific T cells and percentage of IFN-γ production presented as the mean ± SEM (3-4 mice/group). Data are representative of 3 independent experiments. CD8 pmel-1 recovery and IFN-γ secretion were both statistically significantly enhanced when CTX (♦) or CTX and CD8-DC (□) treatment preceded immunization (P = .05).

Figure 6

Local reconstitution of CD8⁺ DCs inhibits systemic concomitant immunity. Mice were inoculated with B16 cells (10⁵) in the left ear on day 0, and reinoculated with B16 (2.5 × 10⁵) in the right shoulder on day 6. Mice were either untreated (no CTX) of treated with CTX (CTX) 4 days before the primary tumor challenge (A). Mice were treated with CTX as described for panel A and received 10⁶ CD8⁺ DCs in the left ear 1 day before the primary tumor challenge or were treated with CTX and received 10⁶ CD8⁻ DCs in the left ear using the same timeline (B). The left graphs represent growth of primary tumors, the right graphs depict growth of challenge secondary tumors. P values for the primary tumor were not statistically significant except no CTX and CTX + CD8⁻ DCs (P = .002). P values comparing growth curves between no CTX with CTX, CTX + CD8⁺ DCs, and CTX + CD8⁻ CCs were < .001, .031, and < .001, respectively. The P value comparing CTX versus CTX + CD8⁺ DCs was .004. The growth of the second challenge tumor inoculum in naive mice without primary tumor challenge (with or without CTX treatment; C). Data from 1 of 2 independent experiments with similar results are shown.