Neuroprotection and brain cholesterol biosynthesis in Huntington’s disease

Sirtuin 2 (SIRT2) is a member of the sirtuin family of protein deacetylases, enzymes that regulate metabolism and longevity and whose inhibition is neuroprotective in Parkinson’s disease models. The study by Luthi-Carter et al. (1) now suggests the mechanism by which SIRT2 inhibition may also function in Huntington’s disease (HD).

The authors (1) suggested that SIRT2 inhibition may act through reduced sterol/cholesterol biosynthesis to curtail mutant huntingtin toxicity. Unfortunately, the effect of SIRT2 inhibitors (AK1 and AGK2 compounds) on in vivo sterol levels and cholesterol biosynthesis in their nonvertebrates’ models was not shown (figure 1 in ref. 1). Three-week-old primary striatal neuronal cultures carrying the N-terminal fragment of human huntingtin (Htt171-82Q) exhibited increased neuroprotection after exposure to AK1 and AGK2. However, gene expression profiles of cholesterol biosynthetic enzymes were evaluated only in 6-wk-old cultures. Furthermore, increased cell death was observed in 3-wk-old Htt171-82Q neurons, contradicting earlier results (2). Comparison of gene expression profiles (figure 3 in ref. 1) between Htt171-82Q/AK1treated and Htt171-18Q/untreated neurons seems counterintuitive, whereas the Htt171-82Q/AK1treated versus Htt171-82Q/untreated comparison was missing. Of the 14 cholesterol biosynthesis pathway genes affected by SIRT2 inhibitors (figure 3a in ref. 1), all were reduced in WT/AK1treated cells, and seven genes were reduced in Htt171-82Q/AK1treated samples; quantitative real time PCR validation of microarray results was only performed in WT/AK1treated cells (figure S5 in ref. 1). The observed effects are, therefore, independent of cytosine adenosine guanine (CAG) trinucleotide repeat expansion and reflect autonomous drug effects. Hence, it generates confusion to indicate that AK1 and AGK2 treatment significantly rescued striatal neurons from mutant huntingtin toxicity by decreasing cholesterol/sterol biosynthesis.

Unexpectedly, BDNF mRNA, a consolidated HD biomarker (3), was not reduced in Htt171-82Q neurons but was reduced after AK1 treatment (dataset S1 in ref. 1), implying that decreased BDNF mRNA levels may be linked to neuroprotection. However, BDNF addition improved survival in the same system (2).

Increased sterol content was shown in Htt171-82Q neurons (figure 3b in ref. 1), but no variations in the mRNA levels of cholesterol biosynthesis genes were detected (dataset S1 in ref. 1). Sterol accumulation may be because of decreased cholesterol efflux, but no mRNA reduction was seen for ATP-binding cassette transporter 1 (or other related genes)—one of the key mechanisms regulating cholesterol efflux.

Overexpression of SIRT2^wt in Htt171-82Q/AK1treated neurons abrogated neuroprotection (figure 2e in ref. 1), but overexpression of SIRT^wt alone was not described. Expression of dominant negative SIRT2 (SIRT2^H150Y) in Htt171-82Q neurons improved survival (figure 2f in ref. 1) and reduced inclusion numbers (figure 2g in ref. 1) and sterol levels (figure 3d in ref. 1), but no mRNA data were provided for cholesterogenic genes. Tagging SIRT2 with a nuclear export signal increased nuclear translocation of procholesterogenic sterol-responsive element-binding protein 2 (SREBP2), sterol content, and cell death (figures S7 and S6a in ref. 1). Conversely, tagging SIRT2 with a nuclear localization signal reduced sterol content (figure S6c in ref. 1), but the expected neuroprotection was not observed (figure S6b in ref. 1). Overexpression of dominant-negative SREBP2 reduced gene transcription and sterol content (figure S8 c and d in ref. 1) but did not influence the neuroprotection-associated number of inclusions (figures 2 d–g and S8e in ref. 1).

To our view, the complicated nature of the experiments and/or their presentation make interpretation difficult. A more systematic and parallel measurement of NeuN+ cells and mRNA levels of key cholesterogenic enzymes and sterol content in control and mutant neurons at the same time points, before and after SIRT2 modulation, would have more rigorously tested the proposed correlation among SIRT2 inhibition, increased neuroprotection, and decreased cholesterol/sterol biosynthesis/level in HD. This hypothesis also contrasts with the evidence that cholesterol precursors and mRNA levels and sterol/cholesterol content are decreased in several HD models (4, 5), indicating that strategies to increase—rather than decrease—cholesterol/sterol biosynthesis may be beneficial.