Abstract
Background Alpha-synuclein aggregates and accumulation are associated with immune activation and neurodegeneration in Parkinson’s disease. The immune activation is not only dependent on the brain-resident microglial cells but also involves peripheral immune cells, such as mononuclear phagocytes including monocytes and dendritic cells, found in the blood as well as infiltrated into the brain. Understanding the involvement of the peripheral immune component in Parkinson’s disease is essential for the development of immunomodulatory treatment, which might modify disease progression. We aimed to study the profile of circulating mononuclear phagocytes in early- and late-stage Parkinson’s disease by analyzing surface-expressed molecules related to phagocytosis, alpha-synuclein sensing, and tissue-migration.
Methods Multi-color flow cytometry on peripheral mononuclear cells from cross-sectional samples of 80 gender-balance individuals with early- and late-stage sporadic Parkinson’s disease, and 29 controls, as well as longitudinal samples from seven patients and one control. Cells were delineated into natural killer cells, monocyte subtypes, and dendritic cells with cell frequencies and surface marker expressions compared between patients and controls, and correlated with standardized clinical motor and non-motor scores.
Results Overall, we found elevated frequencies and surface levels of markers related to migration (CCR2, CD11b) and phagocytosis (CD163) particularly on the elevated classical and intermediate monocytes in patients with Parkinson’s disease for less than five years. This corresponded to a decrease of non-classical monocytes and dendritic cells. We observed an increased HLA-DR expression late in disease and sexual-dimorphism with TLR-4 expression decreased in women with PD but not in males. The disease-associated immune changes on TLR4, CCR2, and CD11b were correlated with non-motor symptoms such as olfaction or cognition. While many alterations were normalized at late disease stage, other changes remained, such as the increased HLA-DR and CD163 expressions.
Conclusions Our data highlight a role for peripheral CD163+ and migration-competent classical monocytes in Parkinson’s disease. The study further suggests that the peripheral immune system is dynamically altered in Parkinson’s disease stages and directly related to both non-motor symptoms and the sex-bias of the disease.
Background
Parkinson’s disease (PD) has classically been described as a purely central nervous system (CNS) disorder with degeneration of dopaminergic neurons in substantia nigra and accumulated aggregates of alpha-synuclein (α-syn) found in Lewy bodies and neurites in the nervous system (1). It is now acknowledged that the immune system influences neuronal health and disease progression. The immunopathology in PD was settled with numerous studies showing microglia activation in post-mortem brains (2), low-level systemic inflammation (3), the association between anti-inflammatory non-aspirin NSAID treatments and decreased PD risk (4, 5), and the association of PD risk with variants in immune-related genes (6, 7), including HLA-DR, which suggest a significant role for the adaptive immune system in PD. Indeed, an imbalance towards increased (autoimmune) Th1 and IL-17-producing Th17 CD4 T-cells have been reported in people with PD (PwP) (8). Furthermore, T-cells from PwP respond to α-syn-derived peptides early in PD (9, 10). Despite the importance of the innate immune system’s role in activating/priming the adaptive immune cells, there remains a paucity of evidence on how the innate peripheral immune cells are altered in PD. Most research has focused on the yolk sac-derived brain-resident microglial cells. However, peripheral immune cells infiltrate the brain and participate in α-syn-clearance and/or inflammation and neurodegeneration (11). With increasing evidence that PD is not only a CNS disorder but also affects and, in some cases, originates in the periphery (12), more knowledge is required regarding the peripheral innate immune cells including natural killer (NK) cells and the peripheral mononuclear phagocytes (MNPs).
MNPs compose diverse populations of ontogenic, phenotypic, and functionally overlapping monocytes (Mos) and dendritic cells (DCs) (13, 14). Mos are phenotypically heterogeneous cells originating from the bone marrow and released into the bloodstream as non-dividing cells. MNPs have important functions in the blood, but also as a myeloid precursor reservoir that can enter tissue and differentiate into Mo-derived tissue macrophages (Mϕs) or tissue DCs, the latter mostly during inflammation (13, 14). Each day, half of the circulating MNPs leave the bloodstream under steady-state conditions (14). Based on the expression of CD14 (toll-like receptor (TLR) 4 co-receptor) and CD16 ((FCγRII) important for antibody-dependent phagocytosis and cytotoxicity), Mos are typically subdivided into classical (c), intermediate (i), and non-classical (nc) Mos: cMo (CD14high/CD16−), iMo (CD14low/high/CD16+), and ncMo (CD14-/low/CD16++) (15). cMos, and to some degree also iMos (14), migrate to inflamed tissue in a CCR2-dependent manner and are highly phagocytic. In contrast, ncMos are described as blood-endothelia patrolling cells that can migrate to non-inflamed tissue, including the brain, in a CX3CR1-dependent manner (14). In PD models and PwP, infiltration of CCR2+ or CD163+ cells seem to partake in the inflammatory and neurodegenerative processes in the brain (16, 17).
However, experimental data on MNP alterations (including DCs) ex vivo in PD are scarce, and the physiological relevance needs further investigation. To accurately study the profile of peripheral MNPs ex vivo in PD, we used multicolor flow cytometry on a cross-sectional cohort of patients with sporadic PD as well as a small number of longitudinal PD case studies. We report here that PD is associated with significant alterations in the expression of the scavenger receptor CD163.
Furthermore, PwP showed disease-stage-related dynamic alteration in the frequency of MNP subtypes and receptors compared to healthy controls (HCs) (Fig 1): Increased cMos and iMos with corresponding decreases of ncMos and DCs in early-stage PD were restored at late-stage PD. In-depth analysis revealed a particular increase in the phagocytosis- and migration-competent CD163+/CCR2+ and CD11b+ cells in early PD, whereas HLA-DR expression was elevated primarily in late-stage PD. Sex-differences were observed for CCR2, CD11b, as well as TLR4. Lastly, mature NK cells had increased CD16 expression in PD. Some of these changes in the MNP population were correlated with PD non-motor symptoms, supporting an association between peripheral innate immune cells and disease pathology.
Summary of study findings based on flow cytometry on peripheral mononuclear cells (PBMCs) from people with Parkinson’s Disease (PD) for less than 5 years (<5y), or 5 years and above (≥ 5y), and healthy control (HC) subjects. Changes in cell type frequencies (arrows in the cell ratio columns) and surface receptor expressions (drawings on each cell type) are shown for the innate immune cells in the PBMC pool: for the overall population of TLR2+ (inclusion marker) mononuclear phagocytes (MNPs) and uniquely for each subtype of monocytes (classical (cMos), intermediate (iMos), and non-classical ncMos)) and dendritic cells (DCs) and (in brown) the TLR2−/CD56dim mature natural killer (NK) cells. Sex-differences (far right column) were observed for the expression levels of TLR4 and to some degree for CCR2 and CD11b with arrows indicating the direction of the change when comparing PD vs. HC cells for the relevant sex. Illustration was made using BioRender.com.
Materials and methods
Cohorts and sample collection
The study was conducted with samples from a cohort of patients with sporadic PD including and HCs recruited at the Department of Neurodegenerative Disorders, University Hospital Tuebingen, as previously described (18, 19) (overview in Supp.Fig.1). PwP were stratified for early (< 5 years from diagnosis) and late (≥ 5 years from diagnosis) stage PD. PwP were diagnosed according to the UK Brain Bank criteria (20). Age and gender-matched controls were assessed by neurologists to have no signs of neurodegenerative diseases. The study was approved by the local ethics committee (480/2015BO2), with all participants providing written informed consent. Peripheral blood mononuclear cells (PBMCs) were collected, isolated, and stored at the Hertie Neuro-Biobank of the University of Tuebingen and later shipped to Aarhus University, Denmark for further analysis, as previously described (18). The cohort was split into two sub-studies: Cohort#1A: Cross-sectional PBMC samples from 80 PwP and 29 healthy controls (HCs) (Table 1); and Cohort#1B: Longitudinal PBMC samples from seven PwP and one HC (Supp.Table 1). Information on comorbidities and non-dopaminergic pharmacological treatment related to inflammation or neurological symptoms for Cohort#1A is provided in Supp.Table 2.
Twenty-nine healthy control (HC) individuals and 80 people with sporadic Parkinson’s disease (PD) at early (<5 years since diagnosis) or late (≥ 5 years since diagnosis) stage. Mean [with range], standard derivation (SD), and the number of individuals for whom the data were available at sampling time (visit) are shown for: L-dopa equivalent daily dose (LEDD), Unified Parkinson’s Disease Rating Scale three (UPDRS III), the Montreal Cognitive Assessment (MoCA) score, the Mini-Mental State Examination (MMSE) original (O) score, MMSE original + converted from MoCA (MMSE O+C), Beck Depression Inventory II (BDI II), and Sniffin’ Stick 12 olfaction test. * Late PD males (LM) were significantly older than HC males (HM) and Early PD males (EM) at the visit; and had significantly higher LEDD than late PD females (LF). EM had significantly higher BDI-II than early PD females (EF), p<0.05. **** For each sex, the HCs were significantly different from both early- and late-stage PD, p<0.0001. Unpaired t-tests or one-way ANOVA with Tukey’s multiple comparison when appropriate.
Flow cytometry
PBMCs were stained for flow cytometry, as previously described (21). Shortly, after thawing and washing, cells were stained using a LIVE/DEAD™ Fixable Near-IR Dead Cell Stain Kit, (Invitrogen, Thermo Fisher Scientific) in DPBS (Biowest), blocked in human AB serum (Sigma), and surface stained with a combination of nine fluorochrome-conjugated anti-human antibodies (Supp.Table 2). The cells were run on a LSR Fortessa (BD) flow cytometer and data was analyzed using FlowJO V10. For further details and gating strategy, see supplementary information.
Statistical analyses
Frequencies and median fluorescence intensity (MFI) data from FlowJo were plotted in GraphPad Prism V7 for statistical analyses: For all datasets in Cohort#1A Gaussian normality distribution was informed with a D’Agostino-Pearson omnibus normality test. Unpaired t-test (or Mann-Whitney) was used to identify sex-differences. Assessments of group differences were performed using 1-way ANOVA (or Kruskal-Wallis test) with Tukey’s (or Dunn’s) multiple comparisons test. Associations of cell surface markers with clinical data were tested using Pearson (or Spearman) correlations as appropriate with two-tailed p values. Significant correlations were also assessed using simple and multiple linear regressions calculated in JMP. Only associations with significant P values for both the correlation and the linear regression were considered, unless indicated.
For comparison of longitudinal sampling at different months in Cohort#1B, normality was assumed and paired repeated-measures mixed-effect model (REML) with Tukey’s multiple comparisons were conducted. A non-parametric paired t-test (Wilcoxon) was used for comparison between first and last visit in the male patients with longitudinal sampling.
Results
Demographics
Cohort#1A included samples from patients with sporadic PD and HCs with similar sex- and age-distribution (Table 1). The subdivision of the PwP group into early- and late-stage PD, using 5 years from diagnosis as a cutoff, resulted in the late PwP males being significantly older than HC males at the inclusion (t-test p 0.0207). Levodopa equivalent daily dosage (LEDD) was higher in late males than females (Table 1). Another sexual-dimorphism was the higher Beck Depression Inventory (BDI-II) score in males vs. females with early PD, which was not statistically significant in late PD, maybe due to the low sample size for this measurement (Table 1).
Eight individuals provided longitudinal PBMC samples and were analyzed separately as Cohort#1B, which included seven patients and one control. Two to four samples were available per individual, with a timeframe of one to sixteen months from first to last sampling visit (Supp.Table 1).
Unraveling the MNPs in the PBMC pool by TLR2 expression
In flow cytometry, the Mos are often defined using forward- and side scatter (FSC and SSC), before subtyping according to CD14/CD16. In this approach, a miner fraction corresponding to circulating DCs is often ignored (14). Here, CD56 and TLR2 were used as inclusion markers for natural killer (NK) cells and MNPs from the PBMCs, respectively. Thus, MNP (TLR2+) included all Mos and some DCs, as shown by tSNE plot of concatenated samples of live cells (Supp.Fig.2). All CD14+ and CD16+, except the CD16+/CD56+ cells, were TLR2+; thus, including all Mos. A subfraction of the TLR2+ cells were CD14-/dim/CD16−/ HLA-DRbright/CD56−(/CD163+/CCR2+); thus, expected to be part of the DC compartment. Hence, by gating live cells on TLR2 vs. CD56, we could distinguish the (TLR2+) MNPs from the (TLR2−/CD56dim) mature- and (TLR2−/CD56bright) precursor NK cells, as well as the remaining PBMC pool (TLR2−/CD56−) (Supp.Fig.3).
The TLR2+ MNPs were divided into four subpopulations based on CD14/CD16 and HLA-DR expression (with other observed characteristics in brackets): cMos CD14++/CD16−/HLA-DRlow (CD11b+/CCR2bright/CD163bright/TLR2bright), iMos CD14++/CD16+/HLA-DRbright (/CD11b+/CCR2bright/CD163bright/TLR2bright), ncMos CD14dim/CD16++/HLA-DRdim (/CD11b-/dim/CCR2+/CD163dim/-/TLR2bright), and the unclassified MNPs/DCs CD14dim/-/CD16−/HLA-DRbright (/CD11b−/CCR2+/CD163Mixed/TLR2dim) (Supp.Fig.3). The expression of TLR4 were sex-dependent in HCs, thus the analysis for this marker was separated by sex.
In-depth analysis of DCs can discriminate these cells into the highly specialized plasmacytoid (p)DC and the antigen-presenting conventional (c)DC. We used an additional antibody panel on three donors and confirmed that cDCs are TLR2+ and thus, included in our MNP population. The CD303+ pDCs important for type I IFN production (13) were TLR2−, thus not included in our MNP gate (Supp.Fig.4). The cDCs can be subdivided into CD141+ cDC1 responsible for antigen cross-presentation to CD8 T-cells; and CD1c+ cDC2 responsible for antigen presentation and priming of CD4+ Th1 and Th17 cells, which is mediated by activated CD163+(/CD14+) cDC2s (13). We showed that cDC1 cells are TLR2− and that cDC2 cells are TLR2+ and primarily located in the “unclassified monocyte”/DC gate, being CD14-/dim. However, with a smaller cDC2 fraction being CD14++, thus in the gate for cMos (Supp.Fig.4). With our gating, cDC2s constituted ∼5% of the MNPs, with approx. 90% of the “unclassified” CD14−/CD16− MNPs being CD1c+ cDC2s. For simplicity, the TLR2+/CD14low/CD16− are referred onward as DCs.
No differences in the frequency of viable cells in the flow staining were observed between PwP and HC groups (Supp.Fig.5A). However, a negative correlation (p=0.002, r −0.414 [−0.205 to −0.587]) between PBMC viability and the MoCA (Montreal Cognitive Assessment) score was observed for PwP (Supp.Fig.5B) (not analyzed in HC due to few MoCA scores informed (n=5)). This association remained significant when adjusted for age at inclusion (Supp.Table 3). Hence, worse cognition correlated with better PBMC viability.
Mature NK cells in PwP are activated and primed for antibody-dependent cytotoxicity
Peripheral NK cells migrate to the CNS during neurodegeneration (22) and might participate in α-syn-clearance (23) and antibody-mediated cytotoxicity. We separated NK cells into mature-(TLR2−/CD56dim) and precursor-(TLR2−/CD56bright) NK cells, and further identified the activated CD16+ cell in the mature fraction (Fig.2A-C). We observed an increase in CD16 MFI, i.e. upregulation of the antibody receptor FCγRIII expression, on the mature NK cells in PwP vs. HC (Fig.2D). This upregulation suggests NK cell activation and that the cells are more prone to kill senescent cells, e.g. those burdened with α-syn (24).
A) Gate for identification of all mononuclear phagocytes (MNPs) and natural killer (NK) cells divided into mature (CD56dim) and precursor (CD56bright) populations. B) Gating and C) distribution of CD16+ cells within the CD56dim mature NK population in healthy controls (HC), early PD (EP), and late PD (LP) patient groups. D) CD16 median fluorescence intensity (MFI) on CD16+ mature NK cells. E) Identification of MNP subtypes and F) comparison on the distribution of MNP subtypes (¤/¤¤ 2-way ANOVA with multiple comparisons adjust p values); the non-adjusted p values for the increase of classical monocytes (cMos) in EP vs. HC was 0.0463. Frequency of the G) classical, H) intermediate (iMo), I) non-classical monocytes (ncMo), and J) dendritic cells (DC) of the total MNP population. Statistical approach G-J): ** non-parametric 1-way ANOVA, cMos: parametric 1-way ANOVA with multiple comparisons. ¤ p <0.05, **/¤¤. p <0.01. G-J are interrelated based on the gating and were Bonferroni corrected for multiple-comparison, all ** P values remained below the adjusted threshold of 0.0125.
PD stage-related changes in the distribution of the MNP subtypes
The percentage of MNPs in PBMCs was similar for HCs (23.17%) to both early (26.69%) and late PD (24.03%). Nevertheless, within the MNP compartment, the distribution of cell subtypes, cMos, iMos, ncMos, and DCs, was altered (Fig.2E-J). The analysis of the dynamic changes in the subpopulation distribution in each group (2-way ANOVA), showed a significant interaction between the effect of subtypes and disease stages (F (6, 424) = 3.761, P=0.0012), with a significant decrease of the cMos from early to late-stage PD and a mirrored increase of ncMos in late compared to early PD (Fig.2F). When each Mo subtype was individually analyzed (Fig.2G-J), we found a significant increase of the iMos at late-stage PD vs. HC (Fig.2H). The iMo changes were reflected in a significant drop of DCs in MNP in early and late PD vs. HC (Fig.2J). Notably, we observe a significant (p 0.012), although weak (r 0.293), positive correlation between the percentages of DCs in the MNP population with the Sniffin’ sticks 12 olfaction scores in PwP. Hence, fewer DCs were associated with more severe hyposmia (Supp.Fig.5C).
Next, we assessed how PD affected the MNP subtypes’ surface expression of HLA-DR, a molecule responsible for antigen presentation. Both iMos and ncMos showed an increase in the HLA-DR MFI, i.e. expression, in late PD compared to HCs (Fig.3A-B) indicating increased antigen presentation capacity in late PD, which might be related to both aging and/or to disease progression.
Median fluorescence intensity (MFI) of HLA-DR on A) intermediate and B) non-classical monocytes C), and MFI of TLR4 on all mononuclear phagocytes (MNPs) D) correlated with Sniffin’ sticks 12 olfaction scores. E) The percentage of CD11b+ MNPs. F) MFI of the CCR2bright population on all MNPs, G) classical and H) intermediate monocytes. A priori identification of sex-differences led to sex-separation for C, D+G-H). Statistics: non-parametric/parametric one-way ANOVA tests with appropriate multiple comparison: */¤ p <0.05, **/¤¤ p <0.01, ***/¤¤¤ p <0.001. Healthy control (HC), early PD (EP), late PD (LP). Mean with SD, Pearson correlation p and r values, and linear regression equation with 95% confidence intervals are shown.
TLR4 correlates with olfaction defect in PD and was reduced in females with PD
We next analyzed the receptor TLR4, which recognizes LPS and α-syn and facilitates α-syn-clearance (25). Due to antibody lot number-associated variance, only 66 samples could be included in the analysis. The analysis revealed a sex-difference with lower MFI on MNPs in male vs. female HCs, thus, sexes were analyzed separately (Fig.3C). This revealed a reduced expression of TLR4 on the MNPs in females at all PD-stages, but particularly in early PD. Thus, HC females had TLR4 levels higher than PD females and all males +/-PD. This decrease was observed across all MNP subtypes (Supp.Fig.6).
Interestingly, TLR4 expression in both males and females with PD was negatively correlated with olfaction, which was not related to age at visit, disease onset, disease duration, nor LEDD. Thus, higher TLR4 expression on MNPs was associated with worse olfaction (Fig.3D). This was also true for TLR4 on cMos and iMos subtypes that were weakly correlated in both sexes. While in males alone, olfaction was strongly correlated with TLR4 on iMos (p 0.048, r −0.436), ncMos (p 0.039, r − 0.565) and particularly on DCs (p 0.0003, r −0.726) (Supp.Fig.7). Interestingly, in females, but not in males the non-motor symptoms scores of hyposmia (Sniffin’ sticks 12) and cognitive decline (MoCA) were significantly correlated (r 0.444, p 0.008), thus, decreased in parallel (Supp.Fig5D-E).
Sex-dependent changes related to MNP migratory capacity
We also examined the expression of two molecules important for MNP migration: the integrin CD11b (Fig.3E) and the chemokine receptor CCR2 (Fig3F-G). Although the percentage of CD11b+ cells in the MNP compartment was similar among sexes in HC, it showed sex-dependent changes in cells from PwP. While the CD11b+ compartment was increased in males with early PD, this was not seen in females (Fig.3E). Although all MNPs expressed CCR2, a bright fraction was easily distinguished (Supp.Fig.3C), which included all cMos and iMos. This CCR2bright fraction showed higher expression of the receptor (MFI) on the MNPs from PwP (Fig3F), with the CCR2 increases on cMos and iMos subtypes being driven by females with (early) PD (Fig3G-H). Altogether, this increase in CD11b and CCR2 suggests mobilization of the Mos from the bone marrow and to inflamed tissue, particularly in early-stage PD; with CD11b as the driver in male patients, while in females this was associated with CCR2. Notably, this sex-bias in CCR2 expression was not apparent when the overall MNPs were considered; hence, our data reveal a need for a more sophisticated analysis of immune markers.
Interestingly, high percentage of cells being CCR2dim in combination with the absence of CD11b correlated positively with better cognition in late PD (Fig.4A). When adjusting for age, however, this only showed a trend (p 0.063) (Supp.Table 4). Conversely, the percentage of CCR2bright MNPs in late PD had a negative linear relationship with the MoCA (Fig.4B), although the significance was also lost when adjusted for age (p=0.12, Suppl.Table 4), suggesting that aging is essential in this process. Furthermore, in females only, the MFI of CCR2 on both cMos and iMos had a negative correlation with MoCA (Fig.4C-D); for the iMos the linear regression was also significant even with age as a confounder (Supp.Table 4). Overall, changes of monocyte migration markers were associated with cognitive symptoms.
A) In late PD, the fraction of CD11b−/CCR2dim cells of the mononuclear phagocytes (MNPs) correlated positively with the MoCA scores (p=0.063 after correction for age covariance). B) In late PD, MoCA also correlated negatively with the total fraction of CCR2+ MNPs, (one outlier removed using Grubs), although after correction for age as covariant the linear relation lost significance with p=0.12. In all females with PD, MoCA was negatively correlated with MFI of CCR2 on C) classical (cMos) and D) intermediate monocytes (iMos), but only iMos showed also a significant linear regression that remained after adjusting for age. Uncorrected values are shown and (* p value) indicates significance did not remain after correction. For all covariance analyses, see Supplementary Table 4.
Peripheral MNP changes related to CD163 distinguish PwP from HCs
We have previously shown that PwP’s blood immune cells in vitro have reduced ability to downregulate CD163 expression on CD14+ Mos upon exposure to α-syn (18); and that PwP have increased soluble (s)CD163 levels in serum (females only) and in CSF (both sexes), with CSF sCD163 levels associated with α-syn and cognitive deficits (26). Based on these findings, we aimed to analyze the ex vivo CD163 expression on MNPs from PwP. Interestingly, PwP showed higher CD163 expression (MFI) on the CD163+ cells on the three CD163bright MNP compartments: cMos, iMos, and DCs; though only in early PD for the DCs. (Fig.5). Furthermore, we observed an increased frequency of CD163+ MNPs in early PD (Fig.5E-F), although this was not associated with any specific MNP subtypes (Supp.Fig8).
Median fluorescents intensity (MFI) of CD163 on A) classical (cMos), B) intermediate (iMos), C) non-classical monocytes (ncMos), and D) dendritic cells (DCs). E) Fixed gating of CD163 vs. CCR2. F) Frequency of CD163+ MNPs based on the CD163 single- and double-positive mononuclear phagocytes (MNPs). G) Distribution and interaction with 2-way ANOVA of the CD163 and CCR2 single- and double-positive/negative cells; with each compartment plotted in H-K). L) Gating of CCR2+ cells with or without co-expression of CD11b; with the majority of CCR2+cells being CD163+ shown by heat mapping. M) The frequency of the CD11b/CCR2 double-positive cells with the majority also being CD163+ based on gating in L). N) The frequency of CCR2+ but CD11b− cells with the majority also being CD163+, based on gating in L). For simplicity, in this figure, we referred to CCR2 as -/+ populations, but for the MNPs it is in fact dim/bright populations. Adjusted multiple comparisons p values: non-parametric/parametric: */¤ <0.05, **/¤¤ <0.01, ***/¤¤¤ <0.001, ****/¤¤¤¤ <0.0001. Healthy control (HC), early PD (EP), late PD (LP) groups. H-K and M-N are interrelated, thus after Bonferroni correcting the P significance threshold to 0.0125, N remained significant, but H (P=0.0482) and M (P=0.0149) did not.
Since infiltration of CD163+ cells has been observed in PD models and post-mortem brains of Alzheimer’s disease and PD (17, 27, 28), we further investigated the highly phagocytic CD163+ cells with respect to the co-expression of the chemokine receptor CCR2, regardless of the traditional MNP subtyping (Fig5E&G-K). CD163/CCR2 double-positive cells were increased in early PD (Fig5H); this was in contrast to the double-negative population that showed a trending (p=0.08) decrease for this group (Fig5G), while no differences were observed for the single-positive cells (Fig5I-J). The fraction of CD163+/CCR2+ cells was normalized in late PD (Fig5H).
Since the migration to (inflamed) brain tissue is also CD11b-related (29, 30), we examined the expression of CD11b on the CCR2+ cells, of which the majority also are CD163+ (Fig.5L-N). MNPs from early PD have more “triple-positive” cells and fewer CCR2+/CD11b− (CD163+) cells compared to HCs, suggesting a relevant role for CD11b in PD. Taken together, the frequency of CD163+ MNPs are upregulated in PwP with increased CD163 surface expression on all migration-competent MNPs (CD11b+ and/or CCR2+), but not on the endothelia patrolling ncMos.
TLR2 expression in longitudinal case reports (Cohort#1B)
The analysis of the different markers and cell populations in the small number of longitudinal samples (Cohort#1B) revealed a progressive increase in the MFI of the oligomeric α-syn receptor TLR2 on total MNPs (first vs. last visit, male PD), as well as in the cMos alone (Supp.Fig.9A-B). We also found a significant increase from baseline sampling to 9-12 months later for the TLR2 MFI on ncMos (Supp.Fig.9D). These TLR2 MFI changes were not reflected in the cross-sectional samples from Cohort#1A when comparing early and late PD (not shown). Thus, for a full comprehension of the immune response in PD, longitudinal cohorts are required. The longitudinal case study revealed no further significant findings (not shown).
Discussion
This study examined the dynamic changes of PBMC frequencies for MNPs and NK cells and their expression of relevant surface molecules in early- and late-stage PD in a cross-sectional cohort (#1A). Our findings show enrichment of cMos and iMos within the first four years of PD diagnosis. The shift in MNP subtypes during early PD resulted in a mirrored contraction of the ncMos and DCs subpopulations, with fewer DCs correlated with hyposmia. Notably, these alterations are related to the increased frequency and expression level of CD163 as well as the migration markers CCR2 and CD11b, especially in early PD. While increased HLA-DR expression was more pronounced in late PD. CD16 expression on mature NK cells was also increased in PwP. The only downregulated immune marker in PD samples was TLR4, which decreased in females with early PD. Interestingly, the TLR4 expression was inversely associated with olfaction in PwP, especially in males. Additionally, higher MNP viability, a lower percentage of CD11b−/CCR2dim, more CCR2bright MNPs, and increased CCR2 expression in females were all related to worse cognitive scores in PwP. Altogether, we showed that the innate immune blood compartment changes during early PD with increases of cells and proteins related to immune activation, phagocytosis, and extravasation. While some of these changes were maintained at late-stage PD, the majority faded maybe as a sign of exhaustion and aging. The PD-associated MNP changes were correlated with non-motor symptoms and interestingly, some of the changes were sex-dependent, which suggests a differential immune response in females and males during the PD neuroimmune process. Therefore, our study warrants longitudinal measurements stratified by sex for all future studies.
Dynamic changes of the MNPs in PD
The distribution of Mo and/or MNP subtypes have previously been investigated, but with varying results between studies. Two German studies, with 10 PwP and HCs in each, showed no difference in blood MNP subpopulations (31, 32) despite altered Mo gene expression profiles and CSF Mo distributions, respectively. In one of these studies, Mo subdivision was simplified by including the iMos in the ncMos gate (thus all CD16+), which might mask changes in the blood and explain their unexpected finding of increased “ncMos” in the CSF (32). Three other studies also reported no Mo subtype changes (33–35). In contrast, another German study of 14 PwP, which included iMos in the cMos gate, reported an increase of cMos and a decrease of ncMos compared to controls (36). Finally, a recent British study in 41 people with early PD (with no inflammatory co-morbidities), described increased cMos, similar to the pattern in our study was driven by the PD patients with a high-risk of dementia. This increase was reflected by a decrease in iMos and ncMos (37). Notably, none of these previous studies analyzed the DC compartment. The lack of consensus between studies might be related to differences in the cohorts used in PD stages and type, medication, comorbidities (not informed in all), age, sex, and flow-gating strategy (38, 39). Here, as a unique approach, we pre-gated MNPs using TLR2. Patients included were sex-balanced, treated (with few exceptions, Table 1), with no known underlying PD high-risk genetic variants, and with similar prevalence of autoimmune conditions and anti-inflammatory treatment as the HCs (Supp.Table 2). This approach and the division into early and late PD, revealed changes related to PD stages.
Accordingly, we found a pattern of an increased cMos in early PD that decreased significantly in late PD (2-way ANOVA). The expansion of the cMos at early-stage PD supports our observation of increased cMos in patients with idiopathic REM sleep behavioral disorder (iRBD), a proposed prodromal form of PD (21). The cMos expansion in early PD was paralleled by a decrease of ncMos and also of DCs (cDC2s), which confirms prior reports of reduced DCs in PwP compared to HCs (40) or to HCs with cytomegalovirus infection (19).
Changes in innate cells linked to the adaptive PD immune response
We have previously observed increased CD56dim mature NK cells in iRBD patients (21). Increase of NK cells and NK activity has been described in PwP (41, 42). In the present study, PwP showed mature NK cells with increased activation by CD16 expression, which might suggest an enhanced antibody-dependent cellular cytotoxicity in PD. Interestingly, IgG deposition on neurons in close proximity to CD16+ (FcγRIII) lymphocytic cells has been shown in post-mortem brains from PwP (43). Hence, the increased expression of CD16 on NK cells might contribute to the neurodegeneration mediated by the adaptive immune response. However, NK cells can also clear α-syn (44), thus they seem relevant for balancing the neuroprotective and neurodegenerative immune processes.
Further suggesting a role for adaptive immunity in PD, we observed an increased CD163 expression on cDC2, which (together with CD14) marks them as mature and inflammatory; thus, responsible for Th1/Th17 polarization of CD4 T-cells (13). This increase of mature cDC2 in early-stage PD, could be associated with the reported increases in Th1 (45) and Th17 cells in PwP (8). Indeed, Th-cells have been proposed to drive neurodegeneration in PD models, a process where HLA-DR plays a key role (11). HLA-DR presents antigen-peptides to T-cells on innate immune cells, thus linking innate and adaptive immune components. Genetic variants of HLA class II are associated with increased PD risk (46). Increased HLA-DR/MHC-II expression has been found on Mos in the CSF of PwP (47), in post-mortem brains of PwP (48, 49), and in primate and rodent models of PD (50, 51), and misfolded α-syn upregulates MHC-II (52). Correspondingly, we found HLA-DR expression increased on iMos and ncMos mainly in late PD; in agreement with a recent paper reporting no changes in the HLA-DR MFI on Mos in early PD (37). Moreover, we have previously observed a general decreased HLA-DR expression on RBD patients’ MNPs (21). Thus, HLA-DR expression seems to be dynamically modulated in each disease stage (prodromal, early and late).
MNPs from early PD are marked for inflamed tissue migration
We and others have shown that MNP-associated changes seem particularly relevant for cognition in PD (26, 37, 40). Elevated CCR2 expression is a signature of newly-generated immature Mos from the bone marrow, marked for migration to blood and tissue. Increased Mo-precursor cells/cMos expressing CCR2 have been described in PD (53, 54) and increased leukocyte CCR2 mRNA expression is associated with cognitive decline (55). Here, we also showed increased CCR2 expression as well as increased CCR2+/CD11b+ and CCR2+/CD163+ double-positive MNPs in early-PD, in contrast to the CCR2+/CD11b− and CCR2+/CD163− MNP populations. Thus, this complex analysis informed about the characteristics of MNPs expressing multiple markers contributing to PD, rather than CCR2 alone. The current cohort mainly included PwP with no (MoCA>26) or (MoCA 18-26) mild cognitive impairment, only three had mild dementia (MoCA<18). Despite the relatively high MoCA scores in the cohort, we observed that the combined absence of both migration markers was associated with better cognition, and accordingly, an increased percentage of CCR2+ MNPs correlated with reduced cognition in late PD. Furthermore, increased expression of CCR2 was in females with PD also correlated with cognitive impairment, suggesting that monocytes with migratory capacity are associated with cognitive decline in PD (26, 37, 39). Indeed a significant immune response in limbic areas and cortex is associated to dementia in PD (56). However, the relation between peripheral migration competent monocytes and cognitive impairment should also be investigated in a cohort with more significant cognitive impairment.
Immunological sexual-dimorphism related to PD
Sex-differences related to the immune system are well-established (57, 58) with recent evidence of sex-divergent monocyte aging (38). Some sex-differences seem constitutive, such as the higher TLR4 expression observed here in female HCs. Whereas, sexual-dimorphism for CCR2 and CD11b expressions were obvious only in PD. The sex-dependent TLR4 differences might be particularly relevant since TLR4 could exert a protective role by clearing α-syn (44, 59). Therefore, the increased TLR4 expression on MNPs in HC females could contribute to the lower PD-risk for females. However, prolonged sensing of misfolded α-syn by TLR4 also contributes to (chronic) inflammation (60), which might be differently regulated in men and women. Accordingly, we observed that in PwP, higher TLR4 expression correlated with reduced olfaction, driven by the males. The relevance of TLR4 is in agreement with our previous data showing that TLR4 expression correlated positively with midbrain immune activation and decreased dopaminergic putaminal transmission in iRBD (male) patients (21). In the RBD cohort, TLR2 expression was correlated with hyposmia (UPSIT). In the present cohort, we detected no significant increase in TLR2 expression in the cross-sectional analysis (Cohort#1A); however, we did observe increasing TLR2 MFI in the overall MNP population and on cMos and ncMos in our longitudinal case reports (Cohort#1B). Previous results reported for TLRs in other PD cohorts are conflicting (34, 35, 37). Several factors might account for this: namely, the constitutively low TLR4 surface expression and its tightly regulated endocytic transport, which may be influenced during blood purification; the age-dependent TLR downregulation (61); and the unspecific binding of non-recombinant TLR2 antibodies (62). Altogether, this makes TLRs difficult markers to compare across studies. However, with the relevant biological function, sex-related difference (for TLR4), and correlation with hyposmia (TLR4 in this study and TLR2 in iRBD patients (21)), TLR2/4 seem to play a relevant role for PD-pathology in agreement with extensive data in PD studies (reviewed (11)).
TLR activation and other inflammatory events induce shedding of the MNP-specific membrane receptor CD163, forming sCD163 (63). In HCs, plasma sCD163 correlates negatively with CD163 MFI, but not with CD163+ cell counts (64). With our previous finding of increased serum sCD163 levels in females with PD (26), we anticipated a decrease in CD163 MFI on MNPs from females with PD. However, we observed a significant increased CD163 MFI in PwP for both sexes, indicating a PD-associated upregulation of CD163, in parallel to the increased shedding observed in serum (in females) and CSF (both sexes) (26). This seems specific for PD since in multiple sclerosis, increased sCD163 was parallel to decreased CD163 MFI (65). The CD163 upregulation in PD does not seem to achieve significance before diagnosis, as we did not observe increased CD163 expression in iRBD patients (21). Nevertheless, in the iRBD cohort, a higher percentage of CD163+ cells was associated with lower immune activation in the midbrain and preserved dopaminergic neurotransmission in putamen shown by PET imaging (21). This suggests a protective role of the CD163 population at this prodromal stage. Indeed, CD163 expression has previously been associated with alternative activation of MNPs considered of anti-inflammatory character (66), however, much remains to be investigated.
CD163+ MNPs from PwP are tagged for migration to inflamed tissue such as brain
The PD-related upregulation of CD163 and co-expression with CD11b and CCR2, proteins related to adhesion and migration of leucocytes into tissue, suggests CD163+ cell infiltration into the brain or other inflamed areas (67). In fact, infiltrated CD163+ cells have been found in the post-mortem brains of PD models and PwP (17, 27, 68). The infiltration and subsequent activation of the CD163+/CCR2+/CD11b+ cells might lead to sCD163 production, which is in agreement with our prior work showing increased sCD163 in CSF of PD patients (26). sCD163 in CSF was related to CSF-α-syn levels and cognitive decline, thus, particularly associated with CNS neuroinflammatory events; since here we described increased CD163 MFI on PD blood cells for both sexes, but we previously only reported increased serum-sCD163 in females with PD (26). Therefore, the activation/sCD163 shedding seems to occur significantly in CNS after MΦ-infiltration in both sexes during PD. Further suggesting this, the CSF-sCD163 levels in PwP were correlated with markers of angiogenesis (VCAM-1, ICAM-1 (CD11b agonist), and VEGF-D) and chemotaxis (CCL4 and CCL2 (CCR2 ligand)) (26). Altogether, CD163+ monocytes with increased expression of CCR2 and CD11b seems to infiltrate the brain of PwP, differentiate into MΦs, and possibly influencing the inflammatory status of the CNS, and thus the neuronal health. Accordingly, modulation of CD163+ cells using targeted dexamethasone in a 6-OHDA rat PD model resulted in neuroprotection (17). However, the specific role of the CD163 receptor in PD is yet to be determined.
The smaller number of HCs might limit the present study. Additionally, alternative separation based on subtypes or fast-vs. slow-progression (37) might challenge the early/late-division used. Nevertheless, the early/late-division revealed otherwise hidden differences, which could also explain the inconsistency in previous studies of MNPs in PD.
Conclusion
Altogether, our study suggests that the PD immune response differs with disease stages by activation in early PD and immune-exhaustion emerging at late-stage. Our data indicate important dynamic alterations in the innate immune compartment, particularly in the phagocytic CD163+ migration competent CCR2+/CD11b+ cMos, but for all TLR2+ MNPs in general. We observed increased expression of innate immune cells and receptors in early PD, while HLA-DR was upregulated mainly in late PD. We also showed significant alterations related to the adaptive immune system, such as the maturation of DCs, which might lead to Th1/Th17 priming. Additionally, we found increased CD16 expression on mature NK cells, relevant for antibody-dependent cytotoxicity. Furthermore, we observed sex-related TLR4, CCR2, and CD11b differences highlighting an additional complexity of the disease. Interestingly, myeloid associated changes were related to non-motor symptoms of PD: namely the associations of TLRs and DCs with hyposmia plus CCR2 with cognition. Altogether, our findings show dynamic alterations in the peripheral immune system in PD and highlight a need to further study CD163+ MNPs, with a focus on the cMos and DCs’ role in PD.
Data Availability
The data supporting the conclusions are included within the article and its additional file.
Declarations
Ethics approval and consent to participate
The study was approved by the Ethics Committee of the University of Tuebingen (Germany) committee (480/2015BO2). All participants providing written informed consent.
Consent for publication
Not applicable.
Availability of data and materials
The data supporting the conclusions are included within the article and its additional file.
Competing interests
The authors declare that they have no competing interests.
Funding Support
Funding support for the research covered by this article was provided by the Michael J. Fox Foundation (MRR), the Bjarne Saxhof Fund administered through the Danish Parkinson’s Foundation (MRR) and the Danish Council for Independent Research (MRR).
Authors’ contribution
SKN: Study design, flow cohort#1A, data analysis, wrote the first manuscript draft; KF: Flow and analysis cohort#1B; MC: Flow; CS: Coordinated sample selection and collected biobank archive info for Cohort#1; DG: isolated PBMC from blood; KB: Supervised biobank; MRR: Developed the concept and designed the study. MRR & SKN: Interpreted the data and wrote jointly the final version. All authors critically revised the manuscript.
Additional information
Supplementary material is found in the PDF file “Additional file 1” containing all supplementary information on materials and methods, Supplementary Table 1-4, Supplementary Figure 8-9, and appropriate references.
Acknowledgments
We acknowledge the invaluable technical help provided by Gitte Ulbjerg Toft (Department of Biomedicine, Aarhus University). Samples were obtained from the Neuro-Biobank of the University of Tuebingen, Germany (https://www.hih-tuebingen.de/en/about-us/core-facilities/biobank/), which is supported by the local University, the Hertie Institute and the DZNE. Flow cytometry was performed at the FACS Core Facility, Aarhus University, Denmark.
Footnotes
List of abbreviations
- (α-syn)
- alpha-synuclein
- (BDI-II)
- Beck Depression Inventory
- (cDC1)
- conventional DC type 1
- (cDC2)
- conventional DC type 2
- (cMo)
- classical monocyte
- (CNS)
- central nervous system
- (DC)
- dendritic cell
- (HC)
- healthy control
- (iMo)
- intermediate monocyte
- (LEDD)
- Levodopa equivalent daily dosage
- (MoCA)
- Montreal Cognitive Assessment
- (MFI)
- median fluorescence intensity
- (MNP)
- peripheral mononuclear phagocyte
- (Mo)
- monocytes
- (ncMo)
- macrophage (Mϕ, non-classical monocyte
- (NK)
- natural killer
- (PBMCs)
- Peripheral blood mononuclear cells
- (pDCs)
- plasmacytoid
- (PD)
- Parkinson’s disease
- (PwP)
- people with PD
- (REML)
- repeated-measures mixed-effect model
- (TLR)
- toll-like receptor
- (tSNE)
- t-distributed stochastic neighbor embedding