Results

Persistently low biopterin and high neopterin plasma levels in multiple sclerosis patients

Neopterin is used as activity and prognostic marker in inflammatory diseases, cancer, some infections and rheumatoid arthritis (Altindag et al., 1998; Husain et al., 2013; Prat et al., 2008; Sucher et al., 2010). Neopterin is produced in excess if the activity and expression of GCH1 exceeds the capacity of the two downstream enzymes (PTPS and SPR) to convert the GCH1 product, neopterin-3-phosphate (neopterin-3P) into BH4. Excess neopterin in serum or plasma mainly originates from activated immune cells. Overall, MS patients had increased neopterin levels (Fig. 1A), but there was no obvious association with the clinical disease activity (Fig. 1B, 1C). In parallel, biopterin levels were persistently reduced (Fig. 1A-C), but there was no inverse direct association of high neopterin with low biopterin, except in few patients (Fig. 1C). High neopterin but low biopterin suggested excessive consumption of BH4 presumably in activated immune cells. In support, biopterin levels dropped over time in human whole blood assays stimulated with LPS (Fig. 1D), and microarrays of the spinal cord of EAE mice showed increased GCH1 but reduced RNA levels of the recycling enzyme, quinoid dihydropteridine reductase (QDPR) (Suppl. Fig. 1).
The pterin data of MS patients suggested two alternative therapeutic interpretations assuming that the biopterin/neopterin balance regulates immune functions, which is strongly suggested by previous publications (Cronin et al., 2018; Huber et al., 1984; Schoedon et al., 1986). Either the supplementation with BH4 may beneficially restore normal antioxidative capacity of BH4 and iron homeostasis, or supplementation may further stimulate immune cell proliferation and activity and increase the autoimmune attack. To answer this question we used two models of autoimmune encephalomyelitis in mice (EAE), with relapsing remitting (RR-EAE; SJL mice) or primary progressive (PP-EAE; C57BL6 mice) course of the disease.

Sapropterin (BH4) treatment aggravates EAE in mice

Mice were treated with sapropterin (BH4) or DAHP (GCH1-inhibitor) or vehicle perorally starting at the day of immunization. Sapropterin treated mice reached higher maximum scores in SJL RR-EAE mice (Fig. 2A) and in C57BL6 PP-EAE mice (Fig. 2B) and increased the frequency of mice with high scores. DAHP oppositely reduced scores during the first relapse (Fig. 2A). Both drugs had no effects on body weights. The C57BL6 experiment was done with vehicle and sapropterin only.

BH4 treatment mildly increases infiltration of lymphocytes in the spinal cord in EAE mice

The observed increase of EAE scores in sapropterin-treated mice was associated with a higher number of T-cells infiltrating the lumbar spinal cord white matter, which was revealed by FACS analyses (Fig. 3A, B) and immunofluorescence analyses (Fig. 4). Both CD4+ and CD8+ T-cells were increased. FACS results for individual myeloid cell populations were not significant, but immunofluorescence studies suggested more hotspots of cellular invasion, in particular with staining of CD11b (activity marker) and F4/80 (Fig. 4). Overall proliferation of T-cells and myeloid cells in the spleen was however not affected by sapropterin or DAHP treatment as compared to vehicle treated EAE mice (Suppl. Fig. 2), suggesting that oral BH4 medication did not reinforce T-cell proliferation but was permissive for CNS infiltration. We therefore asked how BH4 medication could affect BBB integrity.
We have two major hypotheses (i) high levels of circulating BH4 activate endothelial NOS (NOS3) at the BBB, facilitating BBB breakdown (Wu et al., 2009) or (ii) BH4 leads to changes of lipid homeostasis thereby affecting barrier functions. The second hypothesis is based on its coenzyme function for AGMO (Watschinger et al., 2015) and our previous studies in colitis mice where BH4/DAHP altered LPAs and ceramides (Zschiebsch et al., 2016), both clinically relevant for MS (Kurz et al., 2018; Schiffmann et al., 2012; Schmitz et al., 2017). Because NO likely has dual effects in EAE (Giovannoni et al., 1998; Wu et al., 2009) we opted for the putative lipid-paths. It is of note that eNOS is localized at the BBB in endothelial cells (Thiel et al., 2001) whereas AGMO is strongly expressed in choroidal cells at the brain-to-CSF barrier (Suppl. Fig. 4).

Sapropterin (BH4) treatment leads to increased long-chain ceramides in EAE mice

We reanalyzed our previous microarray data of EAE versus naïve mice (Schmitz et al., 2014) (GEO dataset GSE60847) for search of lipid metabolizing genes. The RNA studies of the lumbar spinal cord revealed previously not recognized deficits in EAE of fatty acid desaturases (FADS and stearoyl CoA desaturases SCD1, SCD2) and fatty acid elongases (ELOVL) (Fig. 5A), but increased glucocerebrosidase (GBA) that degrades glucosylceramides (Fig. 5A). GCH1 was increased as expected as well NOS2 and NOS3 (Suppl. Fig. 1). The loss of desaturases, in particular FADS1, which is involved in FA18:2 and FA18:3 generation (linoleic and linolenic acid), is remarkable in light of the protective preserving effects of these omega-lipids in EAE and/or human MS (Bittner et al., 2013; Bjornevik et al., 2019).
A lipidomic screen of plasma and lymph nodes of EAE mice treated with sapropterin or vehicle revealed lipid alterations in plasma but not in the lymph nodes in sapropterin versus vehicle treated EAE mice (Suppl. Fig. 3). In plasma, polyunsaturated fatty acids, in particular linolenic acid (FA18:3), were reduced in sapropterin treated mice (Fig. 5B, 5C). It is of note, that feeding of mice with linolenic acid preserves the BBB in EAE mice because it restores gating properties of TASK1 two-pore potassium channels (Bittner et al., 2009).
Targeted LC-MS/MS lipidomic analyses were used to assess signaling lipids in more detail (Fig. 5D-F; Suppl. Fig. 3C). Using lipid species of five classes (28 different species) as input, canonical discriminant analysis clearly separated groups and allowed a >95% correct prediction of group membership based on the first two CanDisc scores (Fig. 5D). The clear separation was based mainly on ceramides of different chain lengths. Ceramides were increased in sapropterin treated EAE mice and oppositely regulated in DAHP-treated mice as compared to vehicle (Fig. 5E). Sapropterin also increased the pro-inflammatory endocannabinoid, 2-AG (Fig. 5E right), whereas DAHP reduced ethanolamide endocannabinoids (AEA and OEA). Polar plots give an overview of multiple bioactive lipid species (Fig. 5F) and reveal an increase of ceramides and unsaturated LPAs in sapropterin-treated mice, whereas lipids were mostly normal in DAHP treated mice. Scatter plots of the concentrations of individual mice are shown in Suppl. Fig. 3C. Ceramide homeostasis is crucial for the maintenance of plasma membrane integrity (Silva et al., 2007) and increased levels suggest leaky membranes and dysfunctions of lysosomal breakdown (Rao et al., 2007). Because the alterations were evident in plasma, it is a systemic effect reaching the BBB and likely affecting membrane attachment and invasion of immune cells.