Differential memory enrichment of cytotoxic CD4 T cells in Parkinson's disease patients reactive to α-synuclein

Abstract

Parkinson's disease (PD) is a complex neurodegenerative disease with a largely unknown etiology. Although the loss of dopaminergic neurons in the substantia nigra pars compacta is the pathological hallmark of PD, neuroinflammation also plays a fundamental role in PD pathology. We have previously reported that PD patients have increased frequencies of T cells reactive to peptides from α-synuclein (α-syn). However, not all PD participants respond to α-syn. Furthermore, we have previously found that CD4 T cells from PD participants responding to α-syn (PD_R) are transcriptionally distinct from PD participants not responding to α-syn (PD_NR). To gain further insight into the pathology of PD_R participants, we investigated surface protein expression of 11 proteins whose genes had previously been found to be differentially expressed when comparing PD_R and healthy control participants not responding to α-syn (HC_NR). We found that Cadherin EGF LAG seven-pass G-type receptor 2 (CELSR2) was expressed on a significantly higher proportion of CD4 effector memory T cells (T_EM) in PD_R compared to HC_NR. Single-cell RNA sequencing analysis of cells expressing or not expressing CELSR2 revealed that PD_R participants have elevated frequencies of activated T_EM subsets and an almost complete loss of cytotoxic T_EM cells. Flow cytometry analyses confirmed that Granulysin⁺ CD4 cytotoxic T_EM cells are reduced in PD_R. Taken together, these results provide further insight into the perturbation of T cell subsets in PD_R, and highlights the need for further investigation into the role of Granulysin⁺ CD4 cytotoxic T_EM in PD pathology.

Fig. 1. CELSR2⁺ cells are enriched in CD4 T_EM in PD_R participants.

A Representative gating strategy of how CELSR2 + CD4 T cells were identified in four different memory subpopulations in HC_NR (n = 12) and PD_R (n = 12) participants. The frequency of CELSR2+ cells were identified for each of the four subpopulations. The complete gating strategy is displayed in Supplemental Fig. 1. B Frequency of memory (combined central memory (T_CM): CD45RA⁻CCR7⁺; effector memory (T_EM): CD45RA⁻CCR7⁻; and effector memory T cells re-expressing CD45RA (T_EMRA): CD45RA⁺CCR7) vs. naïve T cells (T_N: CD45RA⁺CCR7⁺) among CELSR2+ cells. Results are represented as median with interquartile range. Mann–Whitney U-test. C Frequency of T cell memory subpopulations (T_CM, T_EM, and T_EMRA) among memory CELSR2⁺ CD4 T cells. Results are represented as median with interquartile range. Two-way ANOVA and Tukey’s multiple comparisons test.

Fig. 2. scRNAseq of CD4 T_EM highlights distribution differences specific to CELSR2 expression and/or PD_R.

A Representative gating strategy for how CELSR2⁺ and CELSR2⁻ T_EM cells were identified and sorted from HC_NR (n = 4) and PD_R (n = 4) participants for subsequent scRNAseq analysis. B UMAP of CD4 T_EM CELSR2⁻ and CESLR2⁺ T cells. Cluster frequencies of Cluster 1—Differentiating T_EM (C), Cluster 4—Cytotoxic T_EM (D), and Cluster 6—Activated T_EM (E) were analyzed and compared between HC_NR and PD_R in CD4 T_EM expressing CELSR2 or not. Results are represented as median with interquartile range. Frequencies were compared using a paired t-test within PD_R or HC_NR samples, and an unpaired t-test to compare frequency between PD_R and HC_NR subgroups.

Fig. 3. Overall expression of CELSR2 does not significantly impact transcriptomics, yet correlates with an enhanced cytotoxic gene signature in Cluster 4—Cytotoxic T_EM in HC_NR.

Volcano plots comparing differential gene expression (DEG) of PD_R CELSR2⁺ vs PD_R CELSR2⁻ in Cluster 1—Differentiating T_EM (A) and HC_NR CELSR2⁺ vs HC_NR CELSR2⁻ in Cluster 4—Cytotoxic T_EM (B). Red points are significantly upregulated in CELSR2⁺ samples and blue points are significantly higher in CELSR2⁻ samples (adjusted p value <0.05 and Log₂ fold change >1). C Visualization of leukocyt-mediated cytotoxicity module score in CELSR2⁺ and CELSR2⁻ cells from HC_NR participants. Data were represented as median, and p value obtained using an unpaired t-test. The score was created using genes both detected in this study and included in the leukocyte-mediated cytotoxicity biological process (GO:0001909). Volcano plots comparing differential gene expression (DEG) of PD_R CELSR2⁺ vs HC_NR CELSR2⁺ cells (D) and PD_R CELSR2⁻ vs HC_NR CELSR2⁻ cells in Cluster 1—Differentiating T_EM (E). Red points are significantly upregulated in PD_R⁺ cells and blue points are significantly higher in HC_NR (adjusted p value <0.05 and Log₂ fold change >1).

Fig. 4. Cytotoxic T cells from PD_R are differentially enriched in T_EM than HC_NR.

A Representative gating strategy for the identification of CD4 T cells expressing both Granzyme B (GzmB) and perforin, or Granulysin, and the further memory cell phenotyping of the cytotoxic cell subsets, in HC_NR (n = 11) and PD_R (n = 8) participants. Complete gating strategy is displayed in Supplemental Fig. 5. The frequencies of cytotoxic CD4 T cells identified by the double expression of GzmB and perforin (B) or Granulysin (C) were analyzed. Frequencies of T_EM (CD45RA⁻CCR7⁻) within GzmB⁺Perforin⁺ (D) and Granulysin⁺ (E) were analyzed. Results are represented as median with interquartile range. Mann–Whitney U-test.