Current research in PMM2-CDG tries to increase the flux of Man-6-P (mannose-6-phosphate) into the glycosylation pathway by pharmaceutical means or by mannose derivatives . Mannose was not considered as a potential therapy for PMM2-CDG in recent years. This first approach of a long-term mannose supplementation (≥ 2 years) in PMM2-CDG provides data to show that mannose is not completely inert in PMM2-CDG and suggests that the underrecognized role of mannose in PMM2-CDG should be reconsidered. Long-term oral mannose supplementation in PMM2-CDG was well tolerated, led to biochemical improvements in the majority of patients and suggests possible clinical improvements (Figure 1, 2, S2).
Whereas mannose supplementation is the standard of care treatment for MPI-CDG resulting in favorable effects on the biochemistry and the clinical outcome [5,7], short-term dietary supplementations in PMM2-CDG with mannose at 100 mg/kg b.w. every 3h over 9 days  or 0,17g/kg b.w. every 3,5h over a period of 6 months , as well as a continuous i.v. mannose infusion of 5,7g/kg b.w. over a period of 3 weeks  failed to show improvements in glycosylation patterns or clinical benefits. Mannose therapy in vivo could fail because of Man-6-P being catabolized by the fully operative MPI, transferring the surplus of mannose to glycolysis . In vitro, the glycosylation deficiency in PMM2-deficient fibroblasts can be restored by supplementing more than 250 µmol/l mannose . Assuming, that mutations in PMM2 are hypomorph and reduce enzyme affinity for Man-6-P , it is possible, that exogenous mannose supplementation in vitro leads to an increased intracellular Man-6-P concentration, that counters the increased Km requirements of deficient PMM2 fibroblasts. This may lead to higher levels of Man-1-P, increasing the deficient GDP-mannose pools and culminating in normalization of glycosylation . Another possibility may be cytosolic mannose being directly converted into Man-1-P by another enzyme or system (not detected yet) .
In vivo, mannose therapy needed a long time to show effects in PMM2-CDG patients. Even in MPI-CDG, the first partial corrections in IEF- and SDS- patterns of serum transferrin needed several to occur after initiation of a mannose therapy with a dose of 100mg/kg three times a day . This cannot be explained by the biological half-life of transferrin (CDT = ~14d; Non-CDT= ~8d) and other glycoproteins (AT III: 3d; Protein C: 6-8h). Effects on the IEF pattern would be expected not later than four weeks. A biochemical correction in PMM2-CDG may take a longer time and higher mannose dosage to show positive effects, in the light of countering the Km- requirement of the attenuated PMM2. Further investigations of mannose pools used for N-glycosylation might explain this time-lag in the future. Ichikawa et al. found that the contribution of exogenous mannose to N-glycosylation is higher than previously thought and that other potential sources of mannose such as mannose salvaged from degraded glycoproteins, glycogen and gluconeogenesis do not make significant contributions . In fibroblasts, increased exogenous mannose (1mM) can completely replace glucose-derived mannose and become the sole source of mannose in N-glycans and also contribute to galactose and N-acetylglucosamine in N-glycans [8,26]. Explanations for the higher contribution of mannose to N-glycans may be specific mannose transporters (GLUT-like mannose transporter, SGLT-5 mannose specific transporter) [27,28]. About one third of mannose found in N-glycans takes detours as it is first converted to Frc-6-P and reconverted to Man-6-P again. Since the transient Frc-6-P derived from Man-6-P does not equilibrate with the total cellular pool of Frc-6-P, another suggestion may be the presence of separate Frc-6-P-pools (Frc-6-PGP/Frc-6-PG) like check-points for glycosylation, glycolysis and gluconeogenesis, generated by the anomeric selectivity of Glc-6-P and Man-6-P metabolizing enzymes (Fig. S1) [8,26]. A preferable ratio of α- and ß- Man-6-P as well as of MPI (ß-Man-6-P anomer specific) and PMM2 (α-Man-6-P anomer specific) might result in a higher efficiency of exogenous mannose use in glycosylation . Substances enhancing the impact of check points of mannose flux to the glycosylation pathway may improve the effect of exogenous mannose supplementation. Which other undefined factors and check points of mannose metabolism  have an influence on the effect of mannose supplementation needs to be further investigated.
It has to be considered, that the glycosylation of transferrin and other glycoproteins may improve in time, with age and the degree of liver involvement [23,24,29]. There are 2 major arguments against spontaneous improvement in this study. First, the patients in this study started mannose therapy at very different ages from 1 year to 27 years of age (Table S1). Nevertheless, they showed a similar improvement with similar kinetics (Fig. 1, S2), suggesting that mannose supplementation, not age, was responsible for glycosylation improvement. Secondly, the significant correction of the hypoglycosylated serum transferrin returned to approximately pretreatment patterns after long-term interruption of mannose supplementation (Fig. 1C). This clearly indicates the biochemical effect of mannose on glycosylation and not an improvement with age.
Some patients stopped mannose therapy. Giving mannose with meals 5 times a day is time consuming for parents and guardians. Patients might experience episodes of flatulence or diarrhea, which can make an already seriously ill child even more uncomfortable. Mannose dosages have to be slowly increased over weeks and mannose has to be given with food in order to decrease these problems.
Repetitive doses of orally ingested mannose can maintain elevated blood mannose levels in PMM2-CDG patients . Healthy probands reached blood mannose levels of more than 200 µmol/l after 1h by supplementing 0.2g mannose/kg b.w. . The patients in this study showed fluctuating blood mannose with an average blood mannose concentration not reaching the level of 250 µmol/l. With 1g/kg b.w. mannose per day, blood mannose levels could not be maintained properly over the daily period and night time and were lower on average than the concentration shown to correct abnormal glycosylation in fibroblasts (250 µmol/l) . If one assumes that an increased mannose concentration counters the Km- requirements of deficient PMM2, a constantly increased blood mannose concentration might be more essential, than peak values in order to constantly provide sufficient mannose donors for N- and C- glycosylation. Despite the positive responses in the majority of patients, these circumstances have certainly limited the effectiveness of mannose. Parenteral application, for instance by subcutaneous infusion, might be an approach to reach steady blood mannose levels during day and night time.
Since the collection of the clinical data in this study was done in everyday clinical practice without a specific protocol and without matching a control group, retrospective analysis necessarily introduces some bias, which affect the interpretation regarding outcome and coherence negatively. The findings prove a biochemical effect and suggest a clinical effect of mannose therapy. Current literature gives no indication for a normalization of nerve conduction velocity in patients with PMM2-CDG. This study observed PMM2-CDG patients with mannose therapy developing a normalization of their motor nerve conduction velocity and regaining a redeemable knee jerk. Parents and caretakers uniformly reported improved reactivity, attention and better general state while supplementing mannose. The clinical data in this study give the impression that responders with higher initial tetrasialo-transferrin values (Figure S4) and milder phenotype tended to show an improved clinical development. We could not find a significant relation between a certain genotype and a response to mannose therapy.
No therapy will reverse all symptoms of the disease. There are crucial developmental steps during pregnancy and early childhood being disrupted and leading to evident, irreversible malformations and abnormalities (similar to MPI-patients with ductal plate malformations in the liver ). Different organ manifestations of PMM2-CDG may have differently effective responses to a mannose treatment, due to crucial developmental steps during embryogenesis and infancy being negatively affected by hypoglycosylation . Therefore, very early or even prenatal therapy may be an issue to make a significant difference in improving these children’s health condition.  Detailed clinical efficacy of dietary D-mannose in PMM2-CDG should be tested in a controlled, double blinded, randomized study.