Abstract
The correlation between immunity and Parkinson’s disease was presented in many papers, which also discussed lymphocyte and natural killer cell. But these studies have yielded inconsistent results. To systematically review the relationship between the lymphocyte subsets/natural killer cell and the risk of Parkinson’s disease, we electronically searched the SpringerLink, Web of Science, Ebsco-medline with full text, Pubmed, Elsevier-ScienceDirect, Ovid-lww-oup, Wanfang Data for case-control trials on comparing the number of peripheral blood lymphocyte subsets and natural killer cell in Parkinson’s patients and healthy controls. According to the Cochrane methods, the reviewers selected literature, extracted data, and assessed the quality. Then, a meta-analysis was performed using RevMan 5.2. Finally, 21 case-control trials including 943 cases of Parkinson’s disease were fit into our data analysis. Meta-analysis showed that the decreased numbers of CD3+, CD4+ lymphocyte subsets and the increased number of natural killer cell were found in Parkinson’s disease patients. In the intermediate and late stage of PD, CD8+ lymphocyte subsets had a significant decrement. However, the number of B lymphocyte subsets had no significant association with Parkinson’s disease. The lymphocyte subsets and NK cell may be associated with the risk of Parkinson’s disease.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Parkinson’s disease (PD) is a common degenerative disease, mainly caused by selectively and progressively loss of dopaminergic neurons in striatum and substantia nigra. Age, genetic predisposition, environmental risk factors, and inherent susceptibility of dopaminergic neurons were thought to be the inducements of PD. However, in recent years, studies have shown that inflammation in nerve-system is involved in the progression of PD [1]. In the animal model, someone observed the increment of lymphocyte subsets in the late stage of PD progression [2] and infiltration of CD4+ lymphocytes into the brain [3, 4], which illustrated the lymphocyte subsets are closely associated with Parkinson’s disease. Also several papers reported that natural killer cell (NK cell) was found in central nervous system [5] (CNS). However, these results are not consistent. Therefore, in this study, we carried a case-control study on the correlation of lymphocyte subsets/NK cell and PD by Meta-analysis of Observational Studies in Epidemiology rule [6] to provide new insights of PD pathogenesis.
Methods
Data source, search strategy, and inclusion criteria
SpringerLink, Web of Science, Ebsco-medline with full text, Pubmed, Elsevier-ScienceDirect, Ovid-lww-oup, and Wanfang Data were searched for original articles from inception to March 13, 2017 to identify all case-control studies. The following search items were indicated: (“lymphocyte” or “T cell” or “B cell” or “NK” or “CD4” or “CD8” or “CD3” or “CD19” or “CD56”) and (“Parkinson”). The language of the enrolled studies was restricted to English or Chinese. All the relevant articles were conducted on human subjects, and classified as case-control studies. We also conducted a hand search of the reference lists of the enrolled studies or reviews to identify additional eligible studies. Original trials were eligible for the present meta-analysis if they met the following criteria: (1) any study described the association between lymphocyte subsets/NK cell and Parkinson’s disease risk, (2) studies were case-control or cohort type, and (3) the number of the lymphocyte subsets and NK cell of the cases and controls were available. Studies were excluded if (1) there were no data regarding the associations between lymphocyte subsets/NK cell and PD risk, (2) they were duplicate of previous publications, and (3) they were reviews or abstracts.
Data extraction and study quality assessment
The following data from each article were extracted by two researchers independently: name of first author, year of publication, ethnicity of the participants, average age of cases, disease duration, number of controls and cases, Hoehn and Yahr staging (HY), Unified Parkinson’s disease rating scale III (UPDRS-III) scores, and so on. We extracted the data through the following ways: (1) from the articles directly and (2) from the supplementary materials. The quality of eligible case-control studies was assessed independently using Newcastle-Ottawa Scale [6, 7], which evaluates the studies based on following criteria: (1) was the case definition adequate? (1 point), (2) representativeness of the cases (1 point), (3) selection of controls (1 point), (4) definition of controls (1 point), (5) comparability of cases and controls on the basis of the design or analysis (2 point), (6) ascertainment of exposure (1 point), (7) same method of ascertainment for cases and controls (1 point), and (8) non-response rate (1 point). Studies with a score ≥5 points are considered to be high quality.
Data synthesis, statistical analysis, and heterogeneity
We used mean difference with corresponding 95% CI for continuous outcomes (percentages of each lymphocyte subsets of participants in the trials), then standardized mean difference with a 95% CI for continued outcomes when the unit was not unified (absolute numbers and percentages of cell populations of participants in the trials). All quantitative variables are listed in the form of mean ± SD. We calculated the average mean and SD of multiple groups by the methods described in previous studies [8,9,10]. All the statistical analyses were performed using Review Manager 5.2 (RevMan) from the Cochrane Collaboration. We used the χ2 tests and I 2 statistics to assess the magnitude of heterogeneity. We selected a fixed-effect model if there was no unexplained statistical heterogeneity. If heterogeneity existed, then the random-effect model was used [11]. When I 2 ≥ 50% was observed, we could take some measures to reduce the heterogeneity, such as perform subgroup analysis by age and sensitivity analysis. A two-sided P value <0.05 was considered to be statistically significant.
Results
Search results, study characteristics, and study quality
As presented in Fig. 1, the systematic literature retrieval identified a total of 2046 relevant publications from SpringerLink, Web of Science, Ebsco-medline with full text, Pubmed, Elsevier-ScienceDirect, Ovid-lww-oup, and Wanfang Data. After the titles and abstracts were checked, only 34 articles related to the association between lymphocyte subsets/NK cell and Parkinson’s disease remained. An additional 13 publications were excluded due to the lack of sufficient data. As a result, 21 publications were included in the present meta-analysis [12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32]. Overall, these studies had a relatively high quality judged by the Newcastle-Ottawa Scale (9 scores = 2, 8 scores = 6, 7 scores = 7, 6 scores = 5, 5 scores = 1). The characteristics and quality of the enrolled studies were present in Table 1.
Meta-analysis
The results of meta-analysis were presented in Tables 2 and 3. The decrement of the CD3+ and CD4+ lymphocyte subsets numbers were associated with the Parkinson’s disease risk. In intermediate and the late stage of PD, CD8+ lymphocyte subsets had a significant decrement. On the contrary, the number of NK cell increased in Parkinson’s disease. But no significant relevance was identified between the change of B lymphocyte subsets number and the overall Parkinson’s disease risk.
We observed significant heterogeneity in the analysis for CD3+, CD4+, CD8+, B lymphocyte subsets, and NK cells (P < 0.00001, I 2 = 91%; P < 0.00001, I 2 = 93%; P < 0.00001, I 2 = 93%; P < 0.00001, I 2 = 98%; P < 0.00001, I 2 = 86%, respectively). I 2 from the I 2 test was ≥50%, so a random-effect model was used. To investigate the potential discrepancy, we divided the studies by age, disease duration, HY scores, and UPDRS-III scores into subgroups. In the age <60 group of analysis for CD3+ lymphocyte subsets, age <60 and age 60–65 groups for CD4+, age <65 group for B, and age <65 group for NK cells, heterogeneity decreased (P = 0.15, I 2 = 44%; P = 0.0002, I 2 = 82%; P < 0.0001 I 2 = 82%; P = 0.40, I 2 = 0%; P = 0.04, I 2 = 76%). The trend of changes of WMD/SMD in each subgroup was same as the trend of changes of WMD/SMD in total, except for age >65 group of NK cells, in which NK cells had no significant relevance with PD. And we failed to find any association between ages and CD8+ lymphocyte subset (Table 2).
We also carried out a subgroup analysis based on HY scores. By reading the paper carefully, we found that Gao Ping 2005 [15], Li Fengwu 2012 [20], Tan Feng 2002 [25], Zhao Xu 2013 [29], and Feng Xueyi 2016 [31] divided the PD patients into different subgroups by different HY scores. Unfortunately, Kinya Hisanaga 2001 [18], Lei Gesheng 2001 [19], Li Li 2015 [21], and Qi Jinxing 2004 [23] did not provide the HY scores. So we divided the papers into different HY subgroups by their mean HY scores. Subgroups with different HY scores in above papers were included in different HY subgroups, and papers with no HY scores were included in the “No HY” subgroup. Because the controls in above papers were repeatedly calculated, the total number of controls exceeded the real number. All subjects of the papers provided data for B cell has the mean HY scores in the range of 1 to 3, therefore they cannot be divided into different subgroups. In the HY 1–3 group, HY 3–4 group, HY 4–5 group, and No HY group of analysis for CD3+ lymphocyte subsets, HY 4–5 group for CD4+, HY 3–4, and HY 4–5 group for CD8+, HY 3–4, and HY 4–5 group for NK cells, heterogeneity decreased (P < 0.00001, I 2 = 79%; P = 0.006, I 2 = 72%; P = 0.06, I 2 = 60%; P = 0.2, I 2 = 38%; P = 0.01, I 2 = 74%; P = 0.0004, I 2 = 80%; P = 0.21, I 2 = 34%; P = 0.11, I 2 = 60%; P = 0.26, I 2 = 20%). In the HY 3–4 and HY 4–5 groups for CD8+ subsets, the decreased numbers of lymphocytes had the significant relevance with PD. On the contrary, in the HY 1–3 and HY 3–4 groups for NK cells, we observed no association between numbers of cells and PD. The rest of results of each subgroup is consistent with the total results (Table 3).
Sensitivity analysis and publication bias
By excluding one study at a time to evaluate the influence of an individual study on synthetic statistics, we referred this method as a sensitivity analysis (Table 4). In the subgroup analysis based on age, heterogeneity was still significant in CD3+ forest plots when we excluded FENG Xueyi 2016 [31], but the total heterogeneity and heterogeneity in age >65 group decreased (P < 0.00001, I 2 = 83%; P = 0.0002, I 2 = 78%). The WMDs of age <60 group and age >65 group were higher than that of age 60–65 group. When Claire H.Stevens 2012 [13] in the analysis for CD4+ lymphocyte subset was excluded, the total heterogeneity and heterogeneity in age >65 group decreased too (P < 0.00001, I 2 = 82%; P = 0.0002, I 2 = 74%). Correspondingly, the standardized mean differences made some changes. Like the WMDs of CD3+ analysis, the SMD of age 60–65 group was the lowest. When we excluded the Claire H.Stevens 2012 [13] in the B cell analysis, the total heterogeneity and the heterogeneity in age >65 group reduced (P = 0.11, I 2 = 50%; P = 0.03, I 2 = 80%). In the analysis for NK cells, the exclusion of Natalia Pessoa Rocha 2017 [32] made the total heterogeneity and the heterogeneity in age >65 reduced and disappeared (P = 0.31, I 2 = 16%; P = 0.81, I 2 = 0%). The SMD of age <65 is lower than that of age >65. However, the result in CD8+ did not alter obviously.
In the subgroup analysis based on HY scores, excluding one study at a time did not make the results in CD3+ change obviously. And the higher HY scores a group got, the lower WMD it had. The result in CD8+ is the same as that in CD3+. When we excluded Claire H.Stevens 2012 [13] in the CD4+ analysis, the total heterogeneity and the heterogeneity in HY 1–3 group decreased (P < 0.00001, I 2 = 87%; P < 0.0001, I 2 = 84%). Except for No HY group, the higher HY scores a group got, the lower SMD it had. The exclusion of Natalia Pessoa Pocha 2017 [32] made the total heterogeneity, and the heterogeneity in HY 1–3 group decreased (P < 0.00001, I 2 = 79%; P < 0.00001, I 2 = 88%). The number of NK cells in the HY 3–4 group has no significant relevances with PD, and the SMD of this group is the lowest.
In addition, all the enrolled articles were examined by using funnel plot of RevMan. No obvious biases were identified (Fig. 2).
Discussion
Recently, case-control studies indicated that the number of lymphocyte subsets and NK cell altered in PD patients. In the present study, we performed the meta-analysis on a total 21 studies enrolling 943 PD patients. Our data showed that the decreased numbers of CD3+, CD4+ lymphocyte subsets, and the increased number of NK cell may lead to Parkinson’s disease. In the intermediate and late stage of PD (HY 3–5), CD8+ lymphocyte subsets had a significant decrement. Along with the growth of the age, the numbers of CD3+ and CD4+ lymphocyte subsets decreased more. But when age grows further, the decrease slows down. We also found that the higher HY scores a patient got, the lower the numbers of CD3+, CD4+, CD8+ lymphocytes he/she had. And the number of B lymphocyte subsets had no significant association with PD.
It was confirmed that the peripheral blood lymphocyte subsets had complete dopaminergic system. We guessed that changes in CD3+, CD4+ and CD8+ T lymphocyte subsets may be affected by the neuroendocrine-immune regulation network in the central nervous system after inflammatory reaction. Along with the progress of PD, which means PD patients get more HY scores, the lymphocyte subsets get more suppression. On the other hand, with the increment of age, CD3+, CD4+ T lymphocytes also changed in the peripheral blood of PD patients, the number of lymphocyte subsets gradually decreased compared to the normal control group. But with the further increase in age, the number of lymphocyte subsets in PD patients and the control group tend to be similar. Considering the immune function gradually decreased with the growth of age; therefore, lymphocyte subsets of elderly patients with PD decreased more slightly than that of controls. Someone suggested that the numbers of NK cells may be mediated by the ADCC of K cells.
PD incidence has something to do with genetic mutations. About 10–15% patients have positive family history. However, the enrolled studies included PD patients with PD diagnosed according to the UK PD Brain Bank criteria. So, the patients with family history were excluded. All subjects were sporadic PD patients.
We observed the large heterogeneity in the analysis for CD3+, CD4+, and CD8+ lymphocyte subsets and medium heterogeneity for B cell when some studies were excluded. The possible reasons could ascribe to the heterogeneity of the varying age or different stage of PD patients. As PD patients get older or more HY scores, their symptoms get more serious. When we divided articles by age or HY scores, the decrement of heterogeneity demonstrated that. Other reasons may be the clinical diversity of PD. For example, some patients’ main symptom was tremor, others’ main symptom might be the spasm, but all of them were diagnosed as PD. Therefore, heterogeneity would be expected. In the sensitivity analysis for CD3+, the elimination of Feng Xueyi 2016 [31] can reduce the heterogeneity. When we read the paper again, we found that the difference between the cases and the controls were larger than that in any other studies. However, the study design and interventions have no significant differences with others. The reason may be ascribed to the experiment error. Same situation as Claire H. Stevens 2012 [13] and Natalia Pessoa Pocha 2017 [32]: The elimination reduced heterogeneity in the analysis for CD4+, B cell, and NK cells. However, the characteristics have no significant differences either, and the experimental data extracted for the analysis supported the result. The reason may be due to the experiment error too.
Although we have conducted a general retrieval for all eligible studies, there are several drawbacks that should be mentioned. First, the results may lack statistical power because of the limited sample size and limited number of studies enrolled. Second, level of evidence is lower than cohort studies and RCTs as case-control studies. Third, few studies provided UPDRS-III scores. None of them described details about the UPDRS-III scores. Because of the lack of data, we cannot assess such a correlation at the light of non-motor and pre-motor symptoms of PD. Finally, not all studies use the same criteria to evaluate the severity of PD for each patient. Thus, well-designed studies are needed for further study of the actual relevance between lymphocyte subsets/NK cell and PD risk.
References
Olanow CW, Tatton WG (1999) Etiology and pathogenesis of Parkinson’s disease. Annu Rev Neurosci 22:123–144
Pradhan S, Andreasson K (2013) Commentary: progressive inflammation as a contributing factor to early development of Parkinson’s disease. Exp Neurol 241:148–155
Benner EJ, Banerjee R, Reynolds AD, Sherman S, Pisarev VM, Tsiperson V et al (2008) Nitrated alpha-synuclein immunity accelerates degeneration of nigral dopaminergic neurons. PLoS One 3:e1376
Brochard V, Combadiere B, Prigent A, Laouar Y, Perrin A, Beray-Berthat V et al (2009) Infiltration of CD4+ lymphocytes into the brain contributes to neurodegeneration in a mouse model of Parkinson disease. J Clin Invest 119:182–192
Bokor M, Farago A, Garam T, Malatinszky G, Schnabel R (1993) Antibody-dependent cell-mediated cytotoxicity (ADCC) in Parkinson’s disease. J Neurol Sci 115:47–50
Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D et al (2000) Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis of observational studies in epidemiology (MOOSE) group. JAMA 283:2008–2012
Stang A (2010) Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol 25:603–605
Hozo SP, Djulbegovic B, Hozo I (2005) Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol 5:13
Jiang Z, Dai L, Song Z, Li H, Shu M (2013) Association between C-reactive protein and atrial fibrillation recurrence after catheter ablation: a meta-analysis. Clin Cardiol 36:548–554
Li C, Zhang W, Zhou F, Chen C, Zhou L, Li Y et al (2013) Cholesteryl ester transfer protein inhibitors in the treatment of dyslipidemia: a systematic review and meta-analysis. PLoS One 8:e77049
Mantel N, Haenszel W (1959) Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 22:719–748
Yan C, Yuping J, Ping Z (2004) Study on T lymphocyte subsets in Parkinson’s disease. CHINESE JOURNAL OF CLINICAL NEUROSCIENCES 12:377–380
Stevens CH, Rowe D, Morel-Kopp MC, Orr C, Russell T, Ranola M et al (2012) Reduced T helper and B lymphocytes in Parkinson’s disease. J Neuroimmunol 252:95–99
Niwa F, Kuriyama N, Nakagawa M, Imanishi J (2012) Effects of peripheral lymphocyte subpopulations and the clinical correlation with Parkinson’s disease. Geriatr Gerontol Int 12:102–107
Ping G, Chuanzhong Q, Xiangyuan L, Zhuoyou C, Yaxin Z, Zhixiang Z (2005) Clinical study of T lymphocyte and natural kill cell with Parkinson disease. JOURNAL OF CLINICAL MEDICINE IN PRACTICE 9:52–54
Juan H, Shun Y, Biao C, Yaohua L (2014). Analysis of naive, memory, and regulatory T lymphocyte subgroups in peripheral blood of Parkinson’s wu Hospi- tients. Chinese Journal of Neuroimmunology and Neurology, 196–198
Saunders JA, Estes KA, Kosloski LM, Allen HE, Dempsey KM, Torres-Russotto DR et al (2012) CD4+ regulatory and effector/memory T cell subsets profile motor dysfunction in Parkinson’s disease. J NeuroImmune Pharmacol 7:927–938
Hisanaga K, Asagi M, Itoyama Y, Iwasaki Y (2001) Increase in peripheral CD4 bright+ CD8 dull+ T cells in Parkinson disease. Arch Neurol 58:1580–1583
Gesheng L, Jianting M, Zhuyi L (2001) T lymphocyte subgroups and erythrocyte immune function in Parkinson disease patients. CHINESE JOURNAL OF NEUROIMMUNOLOGY AND NEUROLOGY 8:95–97
Fengwu L, Peilan Z, Chuanhai C, Weimin D (2012) Analysis of T lymphocyte subsets in peripheral blood from patients with Parkinson’s disease. Journal of Tianjin Medical University 18:208–211
Li L (2015) The correlation study of peripheral blood CD4+T cells balance and Parkinson's disease
Gruden MA, Sewell RD, Yanamandra K, Davidova TV, Kucheryanu VG, Bocharov EV et al (2011) Immunoprotection against toxic biomarkers is retained during Parkinson’s disease progression. J Neuroimmunol 233:221–227
Jinxing Q, Jianzhong D, Baohua Q (2004) The study of the level of T lymphocyte subsets, TNF and Il-2 in Parkinson’s disease patients. CHINESE JOURNAL OF NEUROIMMUNOLOGY AND NEUROLOGY 11:243
Chiba S, Matsumoto H, Saitoh M, Kasahara M, Matsuya M, Kashiwagi M (1995) A correlation study between serum adenosine deaminase activities and peripheral lymphocyte subsets in Parkinson’s disease. J Neurol Sci 132:170–173
Feng T, Wei G, Tao H, Ruizhu Z (2003) Changes and its clinical significance of natural kill cell and T lymphocyte subsets in Parkinson disease patients. CHINESE JOURNAL OF IMMUNOLOGY 19:138–140
Baba Y, Kuroiwa A, Uitti RJ, Wszolek ZK, Yamada T (2005) Alterations of T lymphocyte populations in Parkinson disease. Parkinsonism Relat Disord 11:493–498
Nengwei Y, Jie L, Xiaojia L, Jianbo Z, Yousong Y, Hongbin S (2012) Changes of T lymphocyte subgroups in peripheral blood of Parkionson’s disease patients. Sichuan Medical Journal 33:1487–1490
Siyun Z, Congcong S, Limin Z, Luan C, Mingshu M, Zhuolin L et al (2014) Clinical analysis of subpopulation of peripheral T and B lymphocytes in Chinese Parkinson’s disease patients. National Medical Journal of China 0:3726–3730
Xu Z (2013) Parkinson’s disease correlation analysis of T lymphocyte subsets and disease
Kustrimovic N, Rasini E, Legnaro M, Bombelli R, Aleksic I, Blandini F, et al. (2016) Dopaminergic Receptors on CD4+ T Naive and Memory Lymphocytes Correlate with Motor Impairment in Patients with Parkinson’s Disease. Scientific Reports, 6
Xueyi F, Yinghui C, Hui Q, Huibi L (2016) Analysis of T lymphocyte subsets in patients with Parkinson’s disease of different age and severity. Shenzhen Journal of Integrated Traditional Chinese and Western Medicine 26:46–47
Rocha NP, Assis F, Scalzo PL, Vieira ÉLM, Barbosa IG, Souza MSD, et al. (2017) Reduced Activated T Lymphocytes (CD4+CD25+) and Plasma Levels of Cytokines in Parkinson’s Disease. Molecular Neurobiology, 1–10
Acknowledgements
We thank anonymous reviewers for revision of this manuscript.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflicts of interest
The authors declare that they have no conflict of interest.
Funding
This work was supported by the grants from the National Natural Science Foundation of China (grant numbers U1503222).
Electronic supplementary material
ESM 1
(XLSX 13 kb)
Rights and permissions
About this article
Cite this article
Jiang, S., Gao, H., Luo, Q. et al. The correlation of lymphocyte subsets, natural killer cell, and Parkinson’s disease: a meta-analysis. Neurol Sci 38, 1373–1380 (2017). https://doi.org/10.1007/s10072-017-2988-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10072-017-2988-4