PD patients were recruited from the movement disorders clinic in Huashan Hospital between February 2014 and May 2017. Subjects who met the following criteria were selected for genetic tests: (1) a clinical diagnosis of PD according to the United Kingdom PD Society Brain Bank criteria (Gibb and Lees, 1988) (before December 2015) or the MDS clinical diagnostic criteria for Parkinson's disease (Postuma et al., 2015) (after November 2015) ; (2) age at onset < 50, or having a family history of parkinsonism. The ethics committee approved the research of Huashan Hospital. After the receipt of informed consent, blood samples and detailed information on each subject were collected. This information included demographic characteristics, clinical data, family records, neurologic examination, Hoehn and Yahr (H&Y) stage, Unified Parkinson’s Disease Rating Scale part III (UPDRS-III) (Goetz et al., 2004), Mini-mental State Examination (MMSE) (Li et al., 2016), Beck Depression Inventory (BDI) (Huang et al., 2016), Sniffin’ Sticks screening 12 test (SS-12) (Huang et al., 2016), Rapid-eye-movement Sleep Behavior Disorder Screen Questionnaire (RBDSQ) (Stiasny-Kolster et al., 2007) and Epworth Sleepiness Score (ESS) (Chen et al., 2002). Motor function was assessed during the wearing-off phase of levodopa, signified by a loss of drug effect.

Genomic DNA was extracted from peripheral blood leukocytes by blood genomic extraction kit (Qiagen) and subjected to target sequencing (HiSeq2500 sequencer, Illumina) for a panel of PD-related genes (Suppl. 1) and multiple ligation-dependent probe amplification (MLPA) by the SALSA MLPA P051-D1/P052-D2 Parkinson probemix kit (MRCHolland). Mean exon coverage was higher than 100 × , with 95% of target base positions having at least 10 × coverage. Variants different from the reference sequences were filtered to remove common polymorphisms (allele frequency > 1%) according to the 1000 Genomes Project (http://www.1000genomes.org/home), In-house database, ESP6500 (evs.gs.washington.edu/EVS/), and ExAC (Exome Aggregation Consortium, http://exac.broadinstitute.org/). The phenotypes of the filtered genes were compared with the clinical characteristics of the corresponding patient, and the inherited modes were considered to exclude irrelevant genes further. SIFT (http://sift.jcvi.org/), Polyphen-2 (http://genetics.bwh.harvard.edu/pph2/), and MutationTaster (http://www.mutationtaster.org/) were used to predict the impact of the variants on gene function. The possibly pathogenic variants were confirmed in the proband and other family members by Sanger sequencing.

WT human TGM6 cDNA (NCBI transcript NM_198994) was amplified by polymerase chain reaction with designed primers (forward nucleotide sequence 5′-CGCAAATGGGCGGTAGGCGTG-3′ and reverse nucleotide sequence 5′-TAGAAGGCACAGTCGAGG-3′). The MT TGM6 constructs were amplified from the WT TGM6. The products were introduced into a pcDNA3.1-EGFP-3 × Flag vector via the 5′-EcoRI and 3′-BamHI restriction sites. Hence, the cells transfected by the recombinant plasmids expressed EDFP-TG6-3 × Flag fusion proteins. All constructs were verified by Sanger sequencing.

A53T-SH-SY5Y cells overexpressing α-synuclein have been constructed before by stably transfecting SH-SY5Y cells with human SNCA A53T cDNA in our laboratory. A53T-SH-SY5Y cells and HEK293 cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco) supplemented with 10% fetal bovine serum (FBS, Gibco) at 37 ℃ in a humidified incubator of 5% CO₂. Mycoplasma testing was done for the cells. When greater than 80% confluent, cells were transiently transfected with the plasmids of empty vector, WT TGM6, or MT TGM6 using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol [DNA (µg): Lipofectamine 2000 (µl) = 1 : 1.5].

A53T-SH-SY5Y cells in 6-well plates were harvested after transfection for 48 hours (h). The cells were washed with PBS and lysed with RIPA lysis buffer (Beyotime) containing 1% protease inhibitor cocktail (Thermo), 1% phosphatase inhibitor (Thermo) and 1% EDTA (Thermo) for each well. The cell lysates were incubated on ice for 5 minutes (min), sonicated, incubated on the ice again for 20 min and centrifuged at 14000 rpm for 15 min at 4 °C. Supernatants were collected to measure the protein concentration using the BCA protein assay kit (Beyotime) and diluted to keep the same concentration among different groups. Protein aliquots were added with SDS-PAGE loading buffer (Beyotime) and boiled at 100 °C for 10 min.

Aliquots of 20 μg protein were electrophoresed through a 12% SDS gel and transferred onto a PVDF membrane (Merck Millipore). The transferred membrane was blocked in TBST with 5% non-fat milk for 1.5 h at room temperature (RT) and incubated with primary antibodies overnight at 4 °C. The next day, after washing three times, the membrane was incubated for 1 h with the secondary antibody at RT. After washing, the membrane was incubated with the ECL reagent (Merck Millipore). The signal was recorded and analyzed by the Bio-Rad imaging system. With GAPDH as control, the relative protein expression was calculated after normalization.

The primary antibodies used in western blotting included mouse monoclonal anti-flag antibody (Abmart), rabbit monoclonal anti-alpha-synuclein antibody (Abcam), mouse monoclonal anti-LAMP2 antibody (Abcam) and rabbit monoclonal anti-LC3B antibody (Abcam). The secondary antibodies were HRP-conjugated horse anti-mouse IgG (CST) and HRP-conjugated goat anti-rabbit IgG (CST). All of the antibodies were diluted in line with the manufacturer’s instructions.

If the results of western blot reveal an increase in LC3B-II, it is essential to differentiate whether it was caused by an increased flux or a block in the pathway. As such, cells were treated with the autophagy inhibitor bafilomycin A1 (50 nM, MCE) for 6 h at the end of the transfection, with DMSO treated as control.

After transfection for 48 h, HEK293 cells overexpressing WT or MT TG6 were washed twice with PBS and added ddH₂O containing 10 mM DTT and 1 mM EDTA. Cells were harvested, lysed through repeated freezing and thawing six times, and centrifuged at 12000 g for 5 min at 4 °C. Supernatants were collected to estimate the protein concentration by the BCA protein assay kit (Beyotime) and diluted to keep the same concentration among different groups. Aliquots of 10 μg protein were used for TG assay using a TG assay kit (Sigma) according to the manufacturer’s protocol. Absorbance at 450 nm was measured using a microplate reader (Thermo).

Before transfection, A53T-SH-SY5Y cells were grown on coverslips pre-coated with poly-D-lysine in 24-well plates for 24 h. The transfection was performed as described above. After 48 h, the cells were fixed with prechilled 4% paraformaldehyde (500 μl/well) for 10 min at RT. Following three washes with PBS, the cells were incubated in PBS supplemented with 0.2% Triton X-100, 2% BSA and the primary antibody (rabbit monoclonal anti-alpha-synuclein filament antibody, 1 : 2000, Abcam) overnight at 4 °C. After washing three times with PBS, the cells were incubated with the secondary antibody (Alexa Fluor 594 goat anti-rabbit IgG, 1 : 2000, Life Technologies) for 1 h at RT in darkness. Then, after being washed three times, the coverslips were mounted with DAPI Fluoromount-G (SouthernBiotech). Images were acquired with the epifluorescence microscope (Olympus BX60). Fluorescence levels were quantified from three randomly selected images for each group and analyzed through ImageJ software (Schneider et al., 2012). With an outline drawn around each cell, the integrated density of selected cells was acquired and adjusted for the background. The mean intensity of selected cells, equal to the ratio of integrated density adjusted for background and area of the selected cell, was calculated.

A53T-SH-SY5Y cells were plated in 96-well plates (1000 cells/well). 24 h later, cells were transfected with plasmids according to the protocol of Lipofectamine 2000. After 48 h of transfection, effects of WT and MT TGM6 on cell proliferation and cytotoxicity were measured using Cell Counting Kit-8 (CCK-8, Dojindo) and LDH Assay Kit-WST (Dojindo), respectively, according to the manufacturers’ instructions. Optical density (OD) values were measured at 450 nm in the CCK-8 test and 490 nm in the LDH assay. Relative cell viability was calculated according to the following equation: relative cell viability = (OD of the experimental group OD of the blank group)/(OD of the mock group OD of the blank group). Relative cytotoxicity was calculated in the same way.

Data were given as mean ± standard deviation (SD). Statistical analysis of clinical information was done using Chi-square test for categorical variables, unpaired Student's t-test for parametric continuous variables, and Mann-Whitney test for nonparametric continuous variables. All experiments were repeated at least three times. One-way ANOVA analyzed experimental data. Post-hoc tests included the Tukey test for the comparison among all groups or the Dunnett method for the comparison of specific groups. Statistical analysis was performed using Graphpad Prism. P < 0.05 was considered statistically significant.

Genetic tests were conducted in 308 PD patients. In 56 of these patients, pathogenic variants accounting for the disease were found. This included 43 patients with homozygous or compound heterozygous variants in PARK2, four patients with PLA2G6 homozygotes or compound heterozygotes, three patients with homozygous variants in PINK1, two patients with LRRK2 variants, two patients with SNCA variants, one patient with homozygous variants in DJ-1, and one patient with a VPS35 mutation. Besides these pathogenic variants, risk factors comprising GBA variants and PARK2 heterozygotes were found in 29 and 13 patients, respectively. The remaining 210 patients were defined as patients with negative genetic findings. In screening all patients, 12 TGM6 variants of low frequency were found in 15 cases. Among these cases, 14 patients belonged to patients with negative findings, and 1 patient, who carried TGM6 c.7 + 1G > T and compound heterozygous deletion of PARK2, belonged to patients with pathogenic variants. Since PARK2 may significantly influence the characteristic of this patient, this patient was not included in the analysis of TGM6 carriers. Notably, two TGM6 variants, c.331C > T (p.R111C) and c.1550T > G (p.L517W), were found in patients characterized by progressive ataxia and were found to affect the function of the protein TG6, such as the transglutaminase activity. The locations of the variants were labeled on the structure of TGM6 (Fig. 1).

Figure 1.

Schematic representation of the TGM6 protein, showing the position of the mutations found in PD patients.

Among the 15 TGM6 carriers we found (14 probands and one brother of the proband in Family A), eight patients had a family history (Fig. 2 Family A-G), with the other seven patients being early-onset without a family history (numbered as EOPD 1-7). All of these TGM6 cases were summarized (Fig. 2 and Table 1). The redundancy of sequencing coverage for these patients is provided in Suppl. 2.

Figure 2.

PD pedigrees with TGM6 mutations. Family A and B were displayed with the pedigree, the results of Sanger sequencing, and PET imaging. There was a marked reduction of DAT uptake in both bilateral putamen and caudate nuclei for each proband (white arrow). Their FDG PET displayed a PD-related pattern with increased metabolism in the putamen, globus pallidus, thalamus, pontine, and cerebellum (red: increased metabolism; blue: decreased metabolism). Family C-G was shown with the pedigree. Black-filled icons indicate PD patients. Grey-filled icons denote possible PD patients. P, proband; het, heterozygous.

The proband in Family A (AII: 6, Fig. 2) was a 64-year-old male who developed resting tremor, rigidity, and bradykinesia at the age of 56, responding well to dopaminergic therapies (pramipexole 0.375 mg/d). No signs of ataxia were found in this patient. His non-motor symptoms included olfactory dysfunction, constipation, and rapid-eye-movement sleep behavior disorder (RBD). His cognition was unimpaired, with an MMSE score of 28. The brother of the proband (AII: 5), a 66-year-old male, developed parkinsonism at 64 years with olfactory impairment, RBD, and depression, but good cognitive performance. The proband accepted the genetic test. The targeted sequencing achieved an average read depth of 329.22 × and a 96.89% breadth of coverage of the target region at a minimum depth of 10 × (Suppl. 2). No pathogenic mutations were detected other than TGM6 c.331C > T (p.R111C). The variant was confirmed in the brother by Sanger sequencing.

The proband in Family B (BIII: 10, Fig. 2) was a 68-year-old male. At the age of 66, he presented with resting tremor, followed one year later by bradykinesia and instability, with a positive response to dopaminergic therapies (carbidopa-levodopa 50/200 mg/d, pramipexole 0.25 mg/d and trihexyphenidyl 2 mg/d), and some non-motor symptoms of constipation and depression. His cognition was normal in the course of the disease, with an MMSE score of 29. There were no signs of ataxia through physical examinations. Within his family, both his cousin (III-3) and niece (IV-2) had been diagnosed with PD according to the United Kingdom PD Society Brain Bank criteria. The genetic result of the proband showed a heterozygous mutation TGM6 c.1550T > G (p.L517W). The mean depth of target sequencing was 278.71 x and the coverage of the target region was 99.74%. The percentage of the target region with mean depth > 10 x was 97.49%.

For both probands in Family A and Family B, dopamine transporter (DAT) imaging with [¹¹C]-2β-carbomethoxy-3β-(4-fluorophenyl) tropane (CFT) and [¹⁸F]-fluorodeoxyglucos (FDG) positron emission tomography (PET) was performed. There was a marked reduction of DAT uptake in both putamen and caudate nuclei bilaterally for each proband. The uptake ratios for [¹¹C]-CFT are displayed in Table 2. The FDG PET of the proband in Family A displayed increased metabolism in the globus pallidus, thalamus, pontine, and cerebellum. The FDG PET of the proband in Family B showed increased metabolism in the globus pallidus, putamen, and thalamus.

TGM6 variants were also found in other families (Fig. 2 Family C-G). Because the probands refused to consent to genetic tests in family members, we report the genetic results of the probands only.

Among the 15 PD patients with TGM6 variants, 10 cases (66.7%) were male. The mean ± SD of age at onset was 48.3 ± 13.9. The primary manifestation of disease in these patients was typical parkinsonism with slow progress. They exhibited a positive response to dopaminergic therapies (initial levodopa equivalent daily dosage ranging from 37.5 mg/d to 400 mg/d) and had several non-motor symptoms. Their PET results agreed with PD. Clinical information was compared between the 15 TGM6 carriers and the 238 patients without pathogenic mutations (Table 3). No differences were found in gender, age, age at onset, disease duration, or non-motor symptoms between groups. However, UPDRS motor scores of TGM6 carriers were significantly lower than those of patients without pathogenic variants, suggesting that patients with PD and TGM6 variants may present with milder symptoms and progress slower.

Among the TGM6 variants of low frequency found in our PD patients, TGM6 p.R111C and p.L517W were appraised as likely pathogenic, based on the standards and guidelines for the interpretation of sequence variants (Table 4) (Richards et al., 2015). Other TGM6 variants were classified as of uncertain significance. TGM6 p.P359L was previously reported in an early-onset PD pedigree by Westenberger et al. (2016). Therefore, the functional differences between TGM6 WT, p.R111C, p.P359L, and p.L517W formed the focus of our further research. To investigate whether TGM6 variants affect the amount of TG6 protein, plasmids of empty vector, TGM6 WT, R111C, P359L, and L517W were respectively transfected into A53T-SH-SY5Y cells. Because the transfected cells expressing TGM6-3 × flag fusion protein, anti-flag antibodies were used in western blots to reflect the expression level of TGM6. After 48 h transfection, western blot analyses revealed that there was no significant difference among the level of the WT and three MT TG6 (Fig. 3A).

Figure 3.

Expression, distribution, and enzymic activity of TGM6 WT and MT. (A) Western blots demonstrating no difference of TG6 levels among TGM6 WT, R111C, P359L and L517W by the Tukey method. (B) Immunofluorescence displaying a predominant cytoplasmic localization of WT and MT TG6. (C) TG assay showing decreased enzymic properties of TGM6 MT by the Dunnett method. Data were represented as mean ± SD (n = 3 per group). *, P < 0.05; **, P < 0.01.

Since TG6 was EGFP-3 × flag-tagged, we could confirm that the transfection efficiency was similar among different groups through observation by fluorescence. In contrast to the mock group, whose EGFP distributed homogeneously in the nucleus and cytoplasm, cells expressing WT and three MT TG6 all displayed a predominantly cytoplasmic localization of TG6 (Fig. 3B).

To assess the enzymatic properties of TGM6 variants, the TG assay was performed in vitro. Setting the result of the mock group as the background value, the OD values detected were used to represent the activity of TG6. Compared to WT, TGM6 R111C, P359L, and L517W all displayed remarkably lower transglutaminase activity (Fig. 3C).

To explore the effect of WT and MT TG6 on cell viability, CCK8 and LDH assays were used to detect cell proliferation and cytotoxicity, respectively. Our results showed that, compared to TGM6 WT, TGM6 R111C and L517W reduced the cell proliferation in A53T-SH-SY5Y cells (Fig. 4A). Meanwhile, the LDH level in the TGM6 WT group was significantly lower than that in the groups of TGM6 R111C (Fig. 4B). Also, overexpressing WT TG6 tends to reduce cell damage in contrast to the mock group (P = 0.076).

Figure 4.

Effect of WT and MT TG6 on cell viability. (A) CCK8 assay to detect cell proliferation. (B) LDH assay to detect cell cytotoxicity. According to two assay results, overexpressing TGM6 R111C and L517W reduced cell viability in contrast with TGM6 WT. Data are represented as mean ± SD (n = 3 per group) and analyzed by the Dunnett method. *, P < 0.05; **, P < 0.01.

The α-synuclein, is an essential protein in the pathogenesis of PD, was reported to be one of the substrates of TGs. Therefore, we explored the influence of TGM6 WT and MT on α-synuclein and its aggregates, using A53T-SH-SY5Y cells overexpressing α-synuclein. After the transfection of WT or MT TGM6 plasmids for 48 h, western blot was conducted to detect the amount of α-synuclein, and immunofluorescence was performed to analyze the impact of TGM6 on α-synuclein filaments. Results revealed that the cells overexpressing WT TG6 showed a decreased amount of α-synuclein compared to the mock group (Fig. 5A). However, this effect was attenuated when cells overexpressed MT TG6 rather than WT. Through immunofluorescence, we found that there was no significant difference in fluorescence intensity of α-synuclein filaments among WT and MT groups (Fig. 5B and 5C). But the expression of α-synuclein filaments within cell nuclei tended to be higher in the group of TGM6 MT in contrast to TGM6 WT.

Figure 5.

Impact of TGM6 on α-synuclein. (A) Impact of TGM6 on α-synuclein expression detected by immunoblotting. Cells overexpressing WT TG6 showed decreased α-synuclein compared to the mock group. This effect was attenuated in the groups of MT TG6. (B-C) Impact of TGM6 on α-synuclein filaments detected by immunofluorescence. TG6 was EGFP-3 × flag-tagged, shown in green. A53T-SH-SY5Y cells were stained for α-synuclein filament (red). Nuclei were stained with DAPI (blue). There was no difference in fluorescence intensity of α-synuclein filaments between the WT group and each other group. However, α-synuclein filament levels tended to be higher within nuclei in the TGM6 MT group compared to TGM6 WT. Data were analyzed using one-way ANOVA and post hoc Dunnett test (n = 3 per group).

We attempted to find a pathway to explain the change in α-synuclein and cell viability. Autophagy is a major protein degradation pathway to maintain cellular homeostasis. Furthermore, hyperactive autophagy may induce apoptosis. We investigated the key proteins in this pathway, including LC3B, whose conversion from LC3B-I to LC3B-II is essential for autophagosome formation; and P62, a receptor of autophagy; and LAMP2, a lysosome-associated membrane protein. Compared to the mock group, the WT group showed markedly higher levels of LC3B-II/I and LAMP2 and tended to have fewer P62 (Fig. 6). Cells were also treated with the autophagy inhibitor bafilomycin A1. The amount of LC3B-II was significantly higher in the presence of the inhibitor, signifying the occurrence of flux (Fig. 6A). There was no significant difference between the WT group and the variant groups in the LC3B-II/I ratio after bafilomycin A1 treatment. The results of western blots revealed elevated autophagy efficiency in the group of TGM6 WT. In the MT TGM6, all of these trends were reduced.

Figure 6.

Expression levels of autophagy-related proteins. Compared to the mock group, the WT group showed markedly higher levels of LC3B-II/I and LAMP2, as well as tending to have less P62. Results from cells treated with bafilomycin A1 demonstrated that an enhanced autophagy flux caused the increase of LC3B-II/LC3B-I in the TGM6 WT group. Data are represented as mean ± SD (n = 3 per group) and analyzed by the Dunnett method. *, P < 0.05; **, P < 0.01; ***, P < 0.001.