Effect of aging on expression of ALP markers in LRRK2 mice
To evaluate whether age-related dysfunction in striatal macroautophagy and CMA machineries is associated with LRRK2 kinase activity, WB analysis was performed in striatal tissues from 3, 12 and 20-month-old G2019S KI, KO, KD and WT mice. No marker changes across genotypes were observed in 3-month-old mice (Fig. S1). Conversely, changes of LC3-I (F3,28=18.03, p < 0.0001; Fig. 1A), LC3-II (F3,28=2.85, p = 0.055, Fig. 1B), p62 (F3,20=5.37, p = 0.0071), p-mTOR (F3,28=12.94, p < 0.0001; Fig. 1E), LAMP2 (F3,28=8.50, p = 0.0004; Fig. 1G) and GAPDH (F3,20=11.69, p = 0.0001; Fig. 1H) were detected in 12-month-old mice. KD mice showed a ⁓2-fold increase in the abundancy of LC3-I and p62, an 80% elevation of LC3-II and a 72% reduction of p-mTOR with respect to controls. This was associated with an increase of CMA markers LAMP2 (+ 46%) and GAPDH (+ 70%). As far as KO and G2019S KI mice were concerned, a 45% increase of LAMP2 levels was only observed. As AMPK is a functional antagonist of mTOR and promotes autophagy in most systems, AMPK levels were investigated in 12-month-old mice (Fig. S2). AMPK levels were significantly affected across genotypes (F3,20=3.62, p = 0.0308), being a 50% elevation observed in KD mice.
WB analysis at 20 months (Fig. 2) showed significant changes of LC3-II (F3,20=5.49, p = 0.0065; Fig. 2B), p62 (F3,20=8.53, p = 0.0008; Fig. 2D), p-mTOR (F3,20=8.91, p = 0.0006; Fig. 2E), mTOR (F3,20=22.28, p < 0.0001; Fig. 2F) and LAMP2 (F3,20=4.93, p = 0.0100; Fig. 2G) across genotypes. Again, KD mice were most affected, showing increases of LC3-II (+ 30%), p62 (+ 80%) and mTOR (+ 95%) levels along with a 50% reduction of p-mTOR levels. At variance with that observed in younger mice, significant elevations of p62 (+ 47%), mTOR (+ 73%) and LAMP2 (+ 30%) levels were found in KO mice. Conversely, no changes in macroautophagy or CMA markers were observed in G2019S KI mice. To investigate whether the increase of mTOR corresponded to an increase of its activity, the phosphorylation levels of the mTOR substrate, S6K1 were investigated (Fig. S3). No changes in p-S6K1 and S6K1 levels were found (Fig. S3B, C). p-AMPK and AMPK levels were also investigated (Fig. S3D, E). Only p-AMPK levels were significantly affected across genotypes (F3,20=6.93, p = 0.0022). Both KO and KD mice showing marked increases (+ 443% and + 300% respectively), being this effect was just above the limit of significance in KD mice (p = 0.059).
Gene expression analysis of autophagy-related markers at 12 months
To investigate whether changes in protein levels correlated with changes in gene expression, RT-qPCR was performed in striatal extracts from 12-month-old mice (Fig. 3). Changes were observed in mRNA levels of MAP1LC3B (F3,28=5.23, p = 0.0054; Fig. 3A), LAMP2 (F3,27=13.44, p < 0.0001; Fig. 3C), mTOR (F3,28=5.39, p = 0.0047; Fig. 3D) and TFEB (F3,28=3.40, p = 0.0315; Fig. 3E) whereas mRNA levels of p62 were unaffected (Fig. 3B). KD mice showed a ⁓20% reduction of mTOR and TFEB mRNA whereas KO mice a 20% reduction of MAP1LC3B mRNA. Conversely, G2019S KI mice appeared largely impaired at transcriptional level due to the significant ⁓20% downregulation of MAP1LC3B, LAMP2, TFEB and mTOR gene transcripts.
Autophagic flux assessment in LRRK2 KD mice at 12 months of age
The lysosomotropic autophagy inhibitor CQ was employed to determine whether the LC3 changes were due to increased or impaired ALP [35]. To validate the administration protocol, CQ (50 mg/Kg, i.p.) [45] was administered daily for three consecutive days in 3-month-old WT mice, and animals were sacrificed 24 hours after the last injection (Fig. S4A). A significant ⁓30% increase was found in LC3-II levels (df = 10, t = 2.88, p = 0.0165) and LC3-II/I ratio (df = 10, t = 2.44, p = 0.0350), which is considered an index of effective blockage of autophagosome-lysosome fusion (Fig. S4A) [35]. CQ treatment did not alter the levels of other macroautophagy or CMA markers, such as p62, LAMP2, p-mTOR and mTOR in 3-month-old WT mice (Fig. S4B). The same protocol was then applied to assess the autophagic flux in 12-month-old WT, G2019S KI, KO and KD mice (Fig. S4C). CQ administration significantly reduced the levels of LC3-I (-60%, df = 10, t = 4.20, p = 0.0018), LC3-II (-80%, df = 10, t = 5.70, p = 0.0002) and p62 (-40%, df = 10, t = 3, p = 0.0065) in KD mice (Fig S4C), and elevated LC3-I levels in WT (+ 60%; df = 10, t = 3, p = 0.0107) and G2019S KI (+ 110%; df = 10, t = 3, p = 0.0110) mice (Fig. S4C). However, the LC3-II /I ratio was significantly reduced by ⁓40% in G2019S KI (df = 10, t = 6.11, p = 0.0001), KO (df = 10, t = 5.22, p = 0.0004), KD (df = 10, t = 4.66, p = 0.0009) and WT (df = 10, t = 10, p = 0.0040) mice, suggesting an impairment of the autophagic flux in all genotypes (Fig. S4C). This was accompanied by a consistent elevation of p-mTOR levels across all CQ-treated LRRK2 mutants, that was most dramatic in KD mice: G2019S KI (+ 150%, df = 10, t = 2.55, p = 0.0288), KO mice (+ 110%, df = 10, t = 3.72, p = 0.0040), KD mice (+ 1630%; df = 10, t = 3.74, p = 0.0038). KD mice also showed a 60% reduction of mTOR levels (df = 10, t = 3.41, p = 0.0066). LAMP2 levels were found 57% elevated only in WT mice after CQ treatment (df = 10, t = 3, p = 0.0096; Fig.S4C). These results, in particular the lack of an increase of LC3-II and LC3-II/I ratio in CQ-treated WT mice, led us to hypothesize that these changes were due to a tardive feedback mechanism occurring in all genotypes as a consequence of the CQ-mediated blockage of the autophagic flux, as previously reported in WT mice [45]. Therefore, CQ administration was replicated anticipating WB analysis at 4 h after the last CQ injection, i.e. when the ALP blockage is still effective. This new administration protocol was performed only in 12-month-old KD mice, i.e. the only genotype showing alterations in LC3-I and LC3-II levels. At this time-point, LC3-I levels were unchanged (Fig. 4A) whereas LC3-II levels (df = 10, t = 5.63, p = 0.0002; Fig. 4B) and the LC3-II/I ratio (df = 10, t = 10.35, p < 0.0001; Fig. 4C) were 2.5-fold elevated in WT mice, likely due to the inhibition of autolysosome formation. Conversely, CQ-treated KD mice exhibited a 30% reduction of LC3-II levels (df = 10, t = 2.31, p = 0.0432) and LC3-II/I ratio (df = 10, t = 2.61, p = 0.0258), indicating a defective autophagic flux (Fig. 4B, C). Consistently, the cargo protein p62 accumulated in CQ-treated WT mice (+ 78%, df = 10, t = 4.30, p = 0.0016) but not in KD mice (Fig. 4D). Conversely, p-mTOR levels were elevated by 2.5-fold in WT mice (df = 10, t = 6.32, p < 0.0001) and by 12-fold in KD mice (df = 10, t = 19.57, p < 0.0001; Fig. 4E) suggesting a marked mTOR activation in both genotypes. This enhancement of mTOR activation was coupled to a 44% reduction in total mTOR levels in KD mice (df = 9, t = 4.18, p = 0.0024; Fig. 4F). LAMP2 levels were elevated by CQ administration in both WT (+ 28%, df = 10, t = 2.60, p = 0.0265) and KD mice (+ 52% df = 10, t = 5.89, p = 0.0002; Fig. 4G).
Changes in GCase activity and GBA1 expression in LRRK2 mice
Since data pointed to a reduced autophagic flux in KD mice, we investigated whether also the activity of lysosomal GCase was affected in LRRK2 mice (Fig. 5). Significant changes were observed both at 3 months (F3,20=8.46, p = 0.0008; Fig. 5A) and 12 months (F3,24=14.97, p < 0.0001; Fig. 5B). When compared to WT controls, KD mice showed an enhancement at both ages (+ 80% and + 32%, respectively) whereas KO mice only at 12 months (+ 47%), although such increase was significant when the comparison was made against 3-month-old G2019S KI mice (p < 0.05). Conversely, no changes of GCase activity were detected in G2019S KI mice. RT-qPCR analysis interrogated whether changes in GCase activity were accompanied by changes of GBA1 transcript levels in 12-month-old LRRK2-mutant mice. No significant change in GBA1 gene expression was detected across genotypes (F3,28=2.45, p = 0.0844; Fig. 5C) although a tendency towards a reduction (15%) was observed in G2019S KI mice.
Subacute pharmacological inhibition of LRRK2 kinase activity
To test whether pharmacological LRRK2 kinase inhibition replicates the ALP impairment associated with the constitutive kinase silencing in vivo, MLi-2, a potent and brain penetrant LRRK2 kinase inhibitor, was administered subacutely for 7 days in 12-month-old WT and G2019S KI mice. Levels and expression of macroautophagy and CMA markers were evaluated 4 h after the last injection. Target engagement was confirmed by WB analysis of pSer1292 LRRK2 levels, a readout of LRRK2 kinase activity [46], in G2019S KI mice. In a pilot experiment, 10 mg/Kg (i.p.) b.i.d. for 7 days was administered, since this protocol was well tolerated in 3-month-old mice (unpublished data). After a few doses, however, lethality was observed in some 12-month-old mice of both genotypes prompting us to lower the dosage to 5 mg/Kg (i.p.), once a day for 7 days. Under these conditions, no signs of distress or mortality were noticed. WB analysis revealed a 42% reduction of pSer1292 LRRK2 levels in G2019S KI mice (Fig. S5), indicating effective targeting of LRRK2 in vivo. MLi-2 caused a 50% elevation of LC3-I in WT mice (df = 10, t = 2.35, p = 0.00406) but not in G2019S KI mice, and no changes in LC3-II or LC3-II/ I ratio in both genotypes (Fig. 6A). p62 levels were elevated in both WT and G2019S KI mice, although the increase was significant only in G2019S KI mice (+ 80%, df = 10, t = 2.52, p = 0.0303) whereas LAMP2 levels remained unaffected (Fig. 6A). Interestingly, p-mTOR was reduced by ⁓32% (df = 10, t = 2.58, p = 0.00272) and mTOR levels enhanced by ⁓30% (df = 10, t = 2.51, p = 0.0306; Fig. 6A) in G2019S KI mice, whereas no changes were observed in WT mice. RT-qPCR analysis showed that MLi-2 increased MAP1LC3B (+ 35%, df = 10, t = 2.62, p = 0.00257), and reduced mTOR (-25%, df = 10, t = 3.66, p = 0.0044) and TFEB (-20%, df = 10, t = 2.71, p = 0.00218) transcripts in WT mice (Fig. 6B). Moreover, MLi-2 reduced mTOR transcripts in G2019S KI mice (-25%, df = 10, t = 3.86, p = 0.0032) and left unaffected p62 and LAMP2 transcript levels in both genotypes (Fig. 6B).
Immunofluorescence analysis of LC3B puncta and pSer129 α-syn in striatum and substantia nigra
We previously reported accumulation of pSer129 α-syn soluble forms in the striatum and substantia nigra (SN) of 12-month-old G2019S KI mice [43]. To investigate whether changes in ALP markers were associated with significant increase of neuronal pSer129 α-syn inclusions, immunofluorescence analysis in the striatum and SN of 12-month-old WT, KD, KO and G2019S KI mice was carried out (Fig. 7). Significant changes in LC3 puncta were observed across genotypes (F3,28=5.40, p = 0.0046). Consistent with the increase in LC3B-I and LC3B-II detected by WB analysis, LC3B puncta were more abundant (+ 78%) in striatal MAP2+ neurons of KD mice compared to WT mice (Fig. 7A, B) whereas no alterations were observed in KO or G2019S KI mice. On the contrary, pSer129 α-syn signal was elevated in striatal MAP2+ neurons of G2019S KI mice (+ 60%) but not KD and KO mice (F3,28=3.68, p = 0.0237) (Fig. 7A, C). LC3B changes were also detected in TH + neurons of SN (F3,29=12.47, p < 0.0001), with KD and KO mice showing an increase (2.5-fold and 2-fold, respectively) and G2019S KI mice no change (Fig. 7D, E). However, no changes in pSer129 α-syn signal were observed in TH + neurons across genotypes (Fig. 7F).