Tyrosine phosphatase STEP₆₁ in human dementia and in animal models with amyloid and tau pathology

Abstract

Synaptic degeneration is a precursor of synaptic and neuronal loss in neurodegenerative diseases such as Alzheimer’s disease (AD) and frontotemporal dementia with tau pathology (FTD-tau), a group of primary tauopathies. A critical role in this degenerative process is assumed by enzymes such as the kinase Fyn and its counterpart, the phosphatase striatal-enriched tyrosine phosphatase 61 (STEP₆₁). Whereas the role of Fyn has been widely explored, less is known about STEP₆₁ that localises to the postsynaptic density (PSD) of glutamatergic neurons. In dementias, synaptic loss is associated with an increased burden of pathological aggregates. Tau pathology is a hallmark of both AD (together with amyloid-β deposition) and FTD-tau. Here, we examined STEP₆₁ and its activity in human and animal brain tissue and observed a correlation between STEP₆₁ and disease progression. In early-stage human AD, an initial increase in the level and activity of STEP₆₁ was observed, which decreased with the loss of the synaptic marker PSD-95; in FTD-tau, there was a reduction in STEP₆₁ and PSD-95 which correlated with clinical diagnosis. In APP23 mice with an amyloid-β pathology, the level and activity of STEP₆₁ were increased in the synaptic fraction compared to wild-type littermates. Similarly, in the K3 mouse model of FTD-tau, which we assessed at two ages compared to wild-type, expression and activity of STEP₆₁ were increased with ageing. Together, these findings suggest that STEP contributes differently to the pathogenic process in AD and FTD-tau, and that its activation may be an early response to a degenerative process.

Alzheimer’s disease (AD) is the most prevalent form of all dementias. This neurodegenerative disorder is characterized by synaptic and neuronal loss in defined brain areas. The disease is further characterized by aggregates of extracellular amyloid-β-containing plaques and intracellular tau-containing tangles. AD shares tau pathology with frontotemporal dementia with tau (FTD-tau) which exists as various subtypes, including Pick’s disease (PiD), corticobasal degeneration (CBD) and progressive supranuclear palsy (PSP) [1]. Whereas neuropsychological assessment is used for clinical diagnosis of AD and FTD-tau, the definitive diagnosis rests on the assessment of the distribution and severity of tau and amyloid-β deposition in the post mortem brain [2, 3]. As the aggregation process is believed to initiate or at least, in part, drive the degenerative process, transgenic mouse models expressing tau or amyloid-β have contributed to our understanding of the degenerative process leading to synaptic dysfunction in dementias [4].

The synapse and its connectivity are known to be modulated by the selective activity of phosphotransferases, with Fyn kinase upregulating synaptic strength, and the striatal-enriched tyrosine phosphatase 61 (STEP₆₁) downregulating signalling and synaptic connectivity [5, 6]. Substrates of STEP₆₁ include the kinases ERK, Pyk2 and Fyn, as well as the glutamate receptors NMDAR and AMPAR [7]. This combined role allows STEP₆₁ to dampen the response to extracellular signals through the ERK cascade as well as directly initiating internalisation of glutamate receptors [7]. STEP₆₁ dysfunction is associated with many neurological disorders including developmental disorders (autism spectrum disorder and Fragile X syndrome), psychiatric conditions (schizophrenia, anxiety-related and depressive disorders), and neurodegenerative diseases (AD and Parkinson’s disease) [8]. Understanding STEP₆₁ may thus be critical for paving the way for new therapies.

In this study, we first examined STEP₆₁ in cases of AD and mild cognitive impairment (MCI), to determine how STEP₆₁ expression and activity are altered over the progression of AD (Additional file 1: Materials and Methods; Table S1). Previous observations of STEP₆₁ in AD reported increased STEP₆₁ activity compared to healthy controls [9]. We, however, did not detect a similar increase between clinically diagnosed human cohorts, as neither the expression level of STEP₆₁ nor its activity changed significantly in MCI or AD compared to cognitively normal controls (Additional file 1: Fig. S1A). Not surprisingly, the clinically diagnosed controls displayed a range of Braak stages, ranging from 0–III/IV. To further investigate this discrepancy and the variability in the clinically diagnosed control group, we established that STEP₆₁ phosphorylation and expression levels were not correlated with post mortem delay (Additional file 1: Fig. S2), and then we sorted the cohort by pathological criteria based on the Braak staging of tau pathology. These comparisons revealed a significantly increased STEP₆₁ activity at Braak stages I–II, compared to stage 0 or stages V/VI, as well as an increased STEP₆₁ expression at stages I–II compared to V/VI (Fig. 1A). This finding indicates that a vacillating behaviour of STEP₆₁ occurs over the course of AD progression and the accumulation of phosphorylated tau and high molecular weight species of tau. Interestingly, we observed a similar pattern of increase followed by a decline in the postsynaptic density protein (PSD-95) expression in the human samples sorted by Braak staging (Fig. 1A; Additional file 1: Fig. S3).

Fig. 1

Modulation of STEP₆₁ in human dementias and representative mouse models. A Immunoblot of RIPA-extracted tissue from the superior frontal cortex of patients diagnosed with AD, MCI and healthy controls and ordered by Braak stages. Of note, the controls in panels A and C are the same samples, only that in A they are grouped according to their Braak stage (with most controls being clinically normal displaying advanced Braak stages upon autopsy), whereas in C all controls are combined given that for FTD-tau (PiD, CBD and PSP) the Braak staging method is not available. Protein levels were normalised to actin. One way ANOVA multiple comparisons correction was performed on each data set; *p < 0.05; active STEP p = 0.0031, total STEP p = 0.0036, PSD-95, p = 0.0027, tau, p = 0.5730. B Immunoblot and quantification of extrasynaptically enriched fractions from 6-month-old APP23 mice (n = 9) and wild-type (WT) littermate controls (n = 9). Quantification of active and total STEP₆₁ normalised to GAPDH; unpaired t-test, active STEP p > 0.5, and total STEP p = 0.0315. C Immunoblot of RIPA-extracted tissue from the superior frontal cortex of FTD-tau (n = 8) ordered by disease and PSP severity (Braak staging) and controls (n = 8). AT8-positive tau was not detectable in FTD-tau patients. Protein levels were normalised for β-actin. One way ANOVA with multiple comparisons correction performed on each data set; active STEP p = 0.0299, total STEP p = 0.0213, PSD-95, p = 0.0043, tau p = 0.0708. D Immunoblot of the RIPA fraction of cortical tissue from 2-month-old K3 mice (n = 8) and WT littermate controls (n = 8). Quantification of total and active STEP₆₁ normalised to GAPDH; unpaired t-test, active STEP p = 0.1764, total STEP p = 0.0523. E Immunoblot of the RIPA fraction of cortical tissue from 5-month-old K3 mice (n = 5) and WT littermates (n = 5) probed for total and active STEP₆₁. Quantification of active and total STEP₆₁ normalised to GAPDH; unpaired t-test, active STEP p = 0.0007, total STEP p = 0.0015 (see Additional file 2: Statistical information)

We next examined the activation and expression of STEP₆₁ in amyloid-depositing APP23 mice, finding no significant change in the total protein level in cortical lysates at 6 months of age. However, whereas STEP₆₁ was virtually excluded from the synaptic fraction, isolation of the extrasynaptic fraction revealed a significant increase of STEP₆₁ levels and activity compared to non-transgenic littermates (Fig. 1B; Additional file 1: Fig. S4). In addition to amyloid-β, we have previously observed a significant increase in tau protein at 6 months of age in the APP23 mouse model [10]. This model does not display synaptic or neuronal loss with the accumulation of amyloid pathology at 24 months of age, as reported by Boncristiano and colleagues [11].

We next analysed STEP₆₁ in tissue from patients clinically diagnosed with different subtypes of FTD-tau (PiD, CBD and PSP, see Additional file 1: Table S1). Here, we used the same controls as in Fig. 1A, but grouped them together given that there are no Braak stages for FTD-tau. Compared to healthy controls, the expression and activity of STEP₆₁ were significantly decreased in FTD-tau tissue (Fig. 1C). This decrease of STEP₆₁ was observed together with a significant decrease in total PSD-95, reflecting synaptic loss in these patients (Fig. 1C; Additional file 1: Fig. S3). In contrast to MCI and AD patients, the analysis of FTD-tau patients by Braak staging did not reveal differences in STEP₆₁ or PSD-95 levels (Additional file 1: Fig. S1B).

Patients suspected of FTD-tau diagnosis present with a range of clinical symptoms, and have complex pathological characterisations with multiple tau- and co-pathologies upon post mortem analysis [12]. Compared to the human pathology, the K3 mouse model exhibits a simpler presentation of tau and its accumulation. We have previously characterised the expression of tau in the cortex of 2- and 5-month-old K3 mice, revealing a significant increase in the expression of tau at 5 months of age [13]. When we examined cortical tissue from these mice, we found no significant changes in the level of STEP₆₁ at 2 months of age (Fig. 1D). However, at 5 months of age, we observed an increase in the expression level and activity of STEP₆₁, suggesting that the higher expression, accumulation, or prolonged presence of tau could increase STEP₆₁ (Fig. 1E). A limitation of our study is that the K3 model cannot recapitulate the later stages of human disease, as the transgenic line has drifted since its generation and the mice experience no neuronal loss with age anymore despite their tau pathology [14]. This discrepancy from human pathophysiology limits the conclusions that can be drawn about human disease from what we see here in the K3 mouse.

Together, our findings indicate that STEP₆₁ responds to pathological insults in a time- and localization-dependent manner. Observations in human patients and animal models suggest that an increase of STEP₆₁ occurs early in the course of disease, before the overt loss of synapses. A decrease in STEP₆₁ expression correlates with significant synaptic (and neuronal) loss as AD progresses. We also found changes in STEP₆₁ and its activity in the synaptic compartment when amyloid-β initially accumulates, suggesting localised synaptic activation upon insult. Contrasting human FTD-tau to the mouse model of inherited tauopathy, the response of STEP₆₁ to pathological tau likely depends on the severity of pathology and the type of tau that is driving dysfunction. Together, our data add to the multifaceted mechanisms of degeneration in human dementia and corresponding transgenic models. Further studies into STEP₆₁ are warranted, to better understand how this enzyme is differentially regulated in primary and secondary tauopathies.