Tolerance to the neuron-specific paraneoplastic HuD antigen

Abstract

Experiments dating back to the 1940's have led to the hypothesis that the brain is an immunologically privileged site, shielding its antigens from immune recognition. The paraneoplastic Hu syndrome provides a powerful paradigm for addressing this hypothesis; it is believed to develop because small cell lung cancers (SCLC) express the neuron-specific Hu protein. This leads to an Hu-specific tumor immune response that can develop into an autoimmune attack against neurons, presumably when immune privilege in the brain is breached. Interestingly, all SCLC express the onconeural HuD antigen, and clinically useful tumor immune responses can be detected in up to 20% of patients, yet the paraneoplastic neurologic syndrome is extremely rare. We found that HuD-specific CD8+ T cells are normally present in the mouse T cell repertoire, but are not expanded upon immunization, although they can be detected after in vitro expansion. In contrast, HuD-specific T cells could be directly activated in HuD null mice, without the need for in vitro expansion. Taken together, these results demonstrate robust tolerance to the neuronal HuD antigen in vivo, and suggest a re-evaluation of the current concept of immune privilege in the brain.

Figure 1. p321 is the immunodominant CD8+ T cell epitope of HuD.

(a) A representative peptide screen of 16 HuD peptides. Individual or duplicate C57BL/6 mice were immunized with a single HuD peptide emulsified in TiterMax adjuvant. 7 days later, CD8+ T cells were harvested from draining lymph nodes and plated in an IFNγ ELISPOT assay (2×10⁵/well) with EL4 cells pulsed with 10 uM cognate or irrelevant peptide (5×10⁴/well). The assay was performed in triplicate. Means are plotted and error bars represent standard deviations of the mean. Positive peptides were re-screened in triplicate mice. (b) 7 peptides (in bold) were identified as potential CD8+ epitopes from the HuD protein sequence. (c) C57BL/6 mice were immunized with AdVHuD plus PTx. 13 days after immunization, splenocytes were divided into 8 in vitro stimulation cultures and stimulated with each of the 7 HuD peptides or βgal p96. CD8+ T cells were purified from stimulation cultures and plated (10⁴ T cells/well) with cognate or irrelevant peptide-pulsed irradiated EL4 cells (5×10⁴/well) in an IFNγ ELISPOT assay. The assay was performed in triplicate. Means are plotted and error bars represent standard deviations of the mean. Data is representative of three experiments. (d) As a control for in vitro priming, C57BL/6 mice were immunized with AdVβgal+PTx and stimulated in vitro with each of the 7 potential HuD epitopes or βgal p96 and assayed for IFNγ secretion as in (c).

Figure 2. Characterization of HuD p321-specific CD8+ T.

(a) 5×10⁶ HuD p321-specific in vitro stimulated CD8+ T cells were adoptively transferred into Rag^−/− mice (n = 2) with 2×10⁶ C57BL/6 DC pulsed with p321. Mice also received PTx and IL-2. Eight days post transfer, mice were injected with CFSE-labeled syngeneic splenocytes pulsed with HuD p321 (CFSE^hi) or βgal p96 (CFSE^lo). A naïve control mouse without transferred CD8+ T cells was injected with CFSE-labeled splenocytes. 6 hours after target injection, splenocytes were analyzed by FACS for in vivo target cell lysis. A representative mouse is shown. Data is representative of two experiments. (b) C57BL/6 mice (n = 2) were used as recipients of adoptively transferred HuD p321-specific CD8+ T cells as in (a). A representative mouse is shown. Data is representative of two experiments. (c) Primary kidney cells from C57BL/6 mice (D^b+/K^b+) or transgenic Bm1 mice (D^b+/K^b−) were irradiated and pulsed with HuD p321 or βgal p96 and used as stimulators in an IFNγ ELIPOST assay (5×10⁴/well) with 3× restimulated HuD p321-specific or βgal p96-specific CD8+ T cells (10⁴/well). The assay was performed in triplicate. Means are plotted and error bars represent standard deviations of the mean. Data is representative of two experiments. (d) C57BL/6 mice were immunized with either AdVHuD or influenza virus or left untreated (2 mice per group). 15 days after immunization, CD8+ T cells were isolated from the spleen and stained directly ex vivo with anti-CD8+ antibody and PE-labeled tetramer. A portion of splenocytes from each mouse was stimulated in vitro with cognate peptide for 7 days. Naïve mice were stimulated with HuD p321. CD8+ T cells from in vitro stimulation cultures were stained with anti-CD8+ antibody and PE-labeled tetramer. Plots are gated on CD8+ T cells. Data is representative of two experiments.

Figure 3. C57BL/6 mice are tolerized to HuD.

(a) C57BL/6 mice were immunized with AdVHuD or AdVβgal+PTx (2 mice per group). 13 days later, CD8+ T cells were isolated from the spleen and plated in an IFNγ ELISPOT assay (2×10⁵/well) with EL4 pulsed with 10 uM peptide (5×10⁴/well). The assay was performed in triplicate. Means are plotted and error bars represent standard deviations of the mean. Data is representative of four experiments. (b) C57BL/6 mice were immunized with AdVHuD−/+PTx (2 mice per group). 13 days later, splenocytes were stimulated in vitro with 0.5 uM HuD p321. On day 7, CD8+ T cells were plated in an IFNγ ELISPOT assay (10⁴/well) with DC pulsed with 10 uM peptide (7×10³/well). The assay was performed in triplicate. Means are plotted and error bars represent standard deviations of the mean Data is representative of two experiments. (c) Individual HuD^+/+ or HuD^−/− mice were immunized with AdVHuD+PTx and used in an IFNγ ELISPOT assay as described in (a). The assay was performed in triplicate. Means are plotted and error bars represent standard deviations of the mean. Data is representative of four experiments. (d) Half of the spleens from mice immunized in (c) were stimulated in vitro with HuD p321. After 7 days, CD8+ T cells were isolate from stimulation cultures and plated in an IFNγ ELISPOT (10⁴/well) with peptide pulsed EL4 cells (5×10⁴/well).

Figure 4. Comparison of HuA p321-specific CD8+ T cells and HuD p321-specific CD8+ T cells.

(a) Sequences of HuD p321 and HuA p321 (b) RMA/S cells were incubated with serial dilutions of peptide and stained for D^b MHC I. HuD p321 and HuA p321 were assayed. The A2.1 epitope of influenza (M1) was used as a negative control. The D^b epitope of influenza (NP) was used as a positive control. (c) C57BL/6 mice were immunized with individual peptides (NP, HuA p321, or HuD p321) in TiterMax adjuvant (2 mice per group). 7 days later, draining lymph node CD8+ T cells were plated in an IFNγ ELISPOT assay (2×10⁵/well) with peptide pulsed EL4 cells (5×10⁴/well). The assay was performed in triplicate. Means are plotted and error bars represent standard deviations of the mean. Data is representative of four experiments.