3.1. Identification of PK as antigen in TS patients’ sera
Sera from four TS patents and four age-matched controls were used to screen tissue lysates from rat brain cortex, hippocampus, cerebellum, brainstem, striatum, and heart muscle by immunoblot analysis (). For each patient, two separate serum samples were compared, one collected prior to and one collected during an exacerbated episode of tics. The interval between the pre-exacerbation visit and the exacerbation visit was 2.5 (±0.4) months (standard deviation). For two of the four TS patients, exacerbation of TS symptoms corresponded to the clear detection of a protein with a molecular weight of approximately 60 kDa as shown for one of these patients in . The 60 kDa protein was detected in all regions of the brain, with highest levels in the striatum, hippocampus, and cortex. Additionally, this protein was detected in rat heart muscle lysate.
Immunoblot Analysis of Rat Tissue Lysates Probed with Serum from an Individual with a Presumptive Diagnosis of PANDAS
To further characterize this potential autoimmune target, brain tissue was fractionated into membrane and cytosolic components (). The 60 kDa protein occurred almost exclusively in the soluble fractions of all brain tissues analyzed. The soluble proteins from acutely dissected striatum were further fractionated by anion exchange liquid chromatography and immunoblotted using the same TS exacerbated serum used in (). The majority of the 60 kDa protein eluted from the column in fraction 5 and produced a distinct ultraviolet absorbance peak (280 nm) on the corresponding chromatogram. The 60 kDa protein was visible as the major Coomassie blue stained protein in an SDS-PAGE gel ().
Immunoblots Using Sera from an Individual with a Presumptive Diagnosis of PANDAS to Isolate the 60 kDa Protein
The 60 kDa protein was analyzed by MS/MS mass spectrometry and identified as the M1 isozyme of the glycolytic enzyme, pyruvate kinase (MR 57,818) (). To confirm the identity of the protein, the membrane from the FPLC fractionation () was re-probed with an antibody specific for PK (). A single 60 kDa protein was detected in the soluble protein extract loaded onto the Mono Q column and in fraction 5, confirming the identity of this protein as PK.
HPLC/MS/MS Analysis of the 60 kDa Digested Protein and Amino Acid Sequence of PK
3.2. Tissue Distribution and Immunhistochemical Analyses of PK in Rodent Brain
The distribution and relative abundance of PK in different regions of the brain and peripheral tissue was evaluated by immunoblot analysis of rat tissue lysates using a polyclonal antibody to PK (). PK was detected in all regions of the brain with the highest levels in the ventral striatum (nucleus accumbens), cortex, and cerebellum. Moderate levels were detected in striatum, brain stem, hippocampus, and hypothalamus, as well as skeletal muscle.
Tissue Distribution of PK in Brain and Peripheral Tissue, and Immunocytochemical Staining for PK in Striatal Tissue
For immunohistochemical analyses, mouse dorsal striatum tissue was co-stained for the striatal-specific protein, DARPP-32 and PK, and images were captured by confocal microscopy (). Coronal sections of striatal tissue showed intense staining for DARPP-32 (red) in somata (1 and 2), dendrites, axons, and terminals of neurons in the caudate and putamen, in agreement with previous immunohistological and immunocytochemical studies of this protein (Bibb et al., 2001
; Bibb et al., 2000
; Ouimet et al., 1984
). PK (green) colocalized with DARPP-32, showing intense staining in all cell types within the striatum. Less intense staining was also observed in the neuronal tracks (white matter of the corpus callosum).
3.3. Immunoreactivity Between PK and the Streptococcal M Proteins
To assess reactivity between PK and proteins associated with streptococcal bacteria, cytoplasmic or extracellular surface protein extracts were prepared from four clinically isolated infectious strains of GABHS. Streptococcal serotypes are determined by the more than 80 antigenically distinct M proteins displayed on the surface of the bacterium, and the use of genetic probes indicate that there are many more as yet uncharacterized (Cunningham, 2000
). Two of the serotypes used in this study included the serotype M18 strain MGAS8232, which was isolated from a rheumatic fever patient (Smoot et al., 2002
), and the serotype M3 strain MGAS315, which was isolated from a patient undergoing toxic streptococcal shock (i.e.,
septicemia) (Musser et al., 1991
). The anti-PK antibody exhibited reactivity to each of the extracts with pronounced increase in detection of proteins extracted from the bacterial extracellular surface (). Interestingly, the strongest reactivity was detected with the extracellular proteins of the pathogenic strain MGAS315. In agreement with these findings, it has previously been shown that TS patients have increased titers against M12 and M19 proteins (Müller et al., 2001
). While PK was not selectively detected, glycolytic enzymes have been shown to either occur on the surface of streptococci or be secreted and contribute to pathogenesis (Fontân et al., 2000
). These data demonstrate that there are epitopes on the surface of strains of streptococcus, which are similar to antigenic epitopes of PK. Immediately below the immunoblot is the control blot using normal goat serum in place of the PK antibody. There is very little if any reactivity detected even at film exposure times 10 fold higher than blots with the primary anti-PK antibody.
Reactivity of Streptococcal Proteins with Anti-PK Antibody
To further evaluate streptococcal surface antigens for the ability to elicit an immune response to pyruvate kinase, rabbit polyclonal antibodies raised against purified preparations of the M5, M6, and M24 proteins were used to immunoblot PK (). Antibodies to all three serotype M proteins reacted with PK in a concentration-dependent manner with the antibody to M5 being the most sensitive. For comparison, enolase, another glycolytic protein, was immunoblotted with the antibodies to the M proteins (). Antibodies to the M proteins showed a much lower level of reactivity to enolase as compared to PK. Identical blots using normal rabbit serum in place of the primary antibody are shown below each panel and labeled Con. No signal was detected in these controls even at film exposure times 30 fold longer than those using the anti-M protein antibodies. To further evaluate reactivity, cytoplasmic lysates from different rat brain regions were immunoblotted with the anti-M protein antibodies. The antibody against M24 (), as well as those against M5 and M6 (data not shown), detected a 60 kDa protein that corresponded to a protein band of the same molecular weight as detected with an anti-PK antibody. Control blots using normal rabbit and normal goat sera in place of anti-M24 and anti-PK antibodies, respectively, are shown in the lower panel. No signal was detected in these blots at film exposures 7 fold greater than blots with the primary antibodies.
3.4. Screening of Patients’ Sera for Reactivity to Pyruvate Kinase
The prevalence of anti-PK antibodies was next assessed by immunoblot analyses in a cohort of patients with TS with or without OCD. Examples of immunoblots of rat striatal cytosolic extract and pure PK with a set of one patient’s sera (pre-exacerbation, exacerbation and post-exacerbation) are shown in the first three panels of . Reactivity to PK was observed in sera obtained during and following an exacerbation of symptoms but only slightly prior to the exacerbation. When the patient’s serum (exacerbation) was preabsorbed with PK prior to the immunoblot analysis, the 60 kDa protein bands in the cytosolic extract and in the purified PK sample were greatly reduced in intensity (TS, ex+pk). Additionally, immunoblots probed with anti-PK antibody showed that the 60 kDa bands in the cytosolic extract and purified PK appear in the same relative positions as the bands detected with the patient’s serum (Anti-pk).
Analysis of Patient and Control Sera for Immunoreactivity to PK
To further explore the association between streptococcal infections and TS/OCD, sera from patients with streptococcal infections and exacerbations of TS or TS/OCD symptoms were compared to controls for reactivity to PK (). Sera from TS or TS/OCD patients without streptococcal infections showed no greater reactivity to PK than age-matched controls. However, reactivity to pyruvate kinase increased 6–7 fold when symptoms of TS or TS/OCD were co-morbid with streptococcal infections compared to the other patients’ sera. In summary, these data indicate that streptococcal infections result in the production of antibodies that react to PK, and could be the result of an immune response to proteins on the surface of the group A streptococcus that are antigenically similar to PK.