Nature of the Interaction of Alpha-D-Mannose and Escherichia coli Bacteria, and Implications for Its Regulatory Classification. A Delphi Panel European Consensus Based on Chemistry and Legal Evidence

1. Introduction

Urinary tract infection (UTI) is an infection of the urinary system that could affect the kidneys, the ureters, the bladder and the urethra. The pathogen that could cause UTIs in the urogenital tract and the bladder is Escherichia coli (E. coli) in approximately 53% of cases. Many people develop a single episode in their life (50% of them are female), and approximately 15% to 25% of adults and children suffer from chronic symptomatic UTIs, namely recurrent, persistent, re-infected or relapsed UTIs. Women are at greater risk of developing a UTI than men. Current therapy and treatments include antibiotics, methenamine hippurate salts, topical oestrogens, urine alkalisers, dietary supplements and lifestyle/behavioural changes. In clinical practice, many patients do not respond to standard antibiotic treatments producing important patient burden and high cost to healthcare systems [1]. Escalating bacterial resistance to traditional antibiotics and limited efforts in developing new antibiotics require the identification of novel therapies.

The idea to “disarm” bacteria instead of killing them is dated around the ‘80s and has driven structural biology studies and clinical research versus diverse small molecule strategies that inhibit the bacterial binding to the urothelial cells [2,3].

Box 1. D-Mannose – chemistry and physiology.

D-Mannose is a natural aldohexose sugar, a C-2 epimer of glucose, i.e. differing from glucose by inversion of one of the five chiral centers of the molecule, precisely that on the carbon atom 2. (Figure 1) [4]. D-mannose is one of the nine monosaccharides (D-glucose, D-galactose, D-mannose, D-xylose, L-fucose, D-glucuronic acid, N-acetyl-D-glucosamine, N-acetyl-D-galactosamine, and N-acetylneuraminic acid) commonly found in animal glycans and in vertebrate glycoconjugates [2].

Figure 1. Chemical structure of D-Mannose compared to D-Glucose.

Figure 1. Chemical structure of D-Mannose compared to D-Glucose.

D-mannose, which has a physiological blood concentration less than one-fiftieth of that of glucose, is physiologically present in the human body. It is primarily synthesized from glucose or derives from endogenous glycoconjugates. Its catabolism occurs via glycolysis after which it is used for energy or incorporated into glycans. D-mannose contributes to the glycoprotein synthesis and it is involved in the immunoregulation [5,6]. Many cell types have mannose-specific receptors; hence, stable blood mannose levels are important for facilitating efficient, constant mannose uptake to different cells.

D-mannose occurs naturally in many plants and fruits, especially cranberries, in relatively small amounts. However, the bioavailability of mannose for glycan synthesis in these dietary sources is poor, therefore, dietary mannose is not considered as a significant source of D-mannose for humans.

D-mannose is mainly targeted for supporting urinary tract health either as a standalone product or combined with cranberry extract or probiotics. Indeed, although D-mannose is a simple sugar, it is not metabolized in humans [5,7,8,9,10]. Pharmacokinetic studies have shown that at least 90% of ingested D-mannose is efficiently absorbed in the upper intestine, and rapidly excreted from the bloodstream. Its plasma half-time ranges from 30 min to some hours. The large amount is excreted unconverted into the urine within 30–60 min; the remainder is excreted within the following 8 h [5]. Excess D-mannose (20–35% of the dose) enters urine from the blood circulation within 60 min, where it has the potential to interact with mannose-sensitive structures of uropathogenic E. coli (UPEC) and further lowering pathogenic effect of the bacterium [2]. No significant increase in glucose blood levels occurs during this time and D-mannose is detectable in the tissues only at trace level.

Finally, D-mannose used in the N-glycosylation and glycerophospholipid anchor synthesis – as it happens in UTI prevention - appears to originate from enzymatic stereospecific interconversion of glucose, not from diet intake

Both in the EU and the United States, alpha-D-mannose is contained in different types of healthcare products, such as food supplements [4] and it was placed on the European market in 2015 as a Class IIa medical device. D-mannose was classified as borderline product from the Borderline and Classification medical devices expert group in 2019 borderline manual [11] excluded D-mannose from the list.

The criterion for establishing the regulatory framework in which a product may fall starts from the definition and identification of the mechanism by which it performs its main action [12]. A very relevant documentation supporting the discussion is the Guidance on Borderline Between Medical Devices and Medicinal Products Under MDR (MDCG 2022–5), in which the definition of medical device derives from the Art. 2 of the 2017/745 MDR regulation [13]. By this definition, for medical devices, the main action for intended use must not act in or on the human body by pharmacological, immunological or metabolic means, although it can be supported by one of these. However, the definitions of pharmacological, immunological or metabolic action are not always unique and the dividing line between mechanism of action and related actions in the organism could be unclear. Describing such a line is the basis of the evaluation of products’ primary characteristics. Considering the quality and intensity of the mechanism of action as well as the intended use can also support when discriminating between intended main action and not intended, accessory actions [12,14].

The consensus panel aimed at discussing the nature of alpha-D-mannose as a medicinal product or medical device in the interaction with E. coli bacteria and it provides additional value in comparison to previous publications thanks to the systematic review approach [4].

2. Methods

2.1. Delphi Panel

The consensus was structured according to the modified Delphi panel method [15].

The expert group was formed representing the key expertise with a specific interest in the topic: key opinion leaders were covering different fields to include in the discussion the point of view of the physicians preventing UTIs in their everyday life (urologists) and of the pharmacologists/biotechnologists for their deep knowledge of the different interactions leading to therapeutic action. The point of view of professionals with thorough expertise in the regulatory process of medicinal products and medical devices was also integrated. The panel was also representative of European countries. The median h-index was 48 (range: 38-74).

A Likert scale of 1–7, with 1 expressing no agreement and 7 total agreement was used. The cut-off for the agreement was set at 6 (5 was considered as expressing uncertainty). Seventy-five per cent (6 out of panellists reporting 6 or 7 scores) was chosen as an appropriate cut-off based [16].

2.2. Prisma research

Prisma 2020 expanded checklist and flow diagram were used as appropriate to systematic reviews of studies that evaluate the effects of health interventions [17]. All studies (in vitro and in vivo) explicitly related to the mechanism of action of D-mannose in the binding of E. coli preventing UTI were included. All abstracts were included. All other articles considered relevant by the panel, as far as the regulatory and pharmacological point of view, guidelines and directives, including papers from references were manually retrieved. Studies without abstracts, studies related to the clinical value of D-mannose, studies related to other pathogens different than E. coli, as well as abstracts written in languages requiring translation (Russian, Bulgarian, Japanese) were excluded.

The PubMed database was searched out on 10th October 2022. MESH research was implemented ("Urinary Tract Infections/drug therapy"[Mesh] OR "Urinary Tract Infections/etiology"[Mesh] OR "Urinary Tract Infections/immunology"[Mesh] OR "Urinary Tract Infections/metabolism"[Mesh] OR "Urinary Tract Infections/microbiology"[Mesh] OR "Urinary Tract Infections/physiology"[Mesh] OR "Urinary Tract Infections/physiopathology"[Mesh] OR "Urinary Tract Infections/prevention and control"[Mesh] OR "Urinary Tract Infections/therapy"[Mesh])) AND "Mannose"[Mesh]). A search without MESH was also implemented ((D-mannose [Title/Abstract]) AND (urinary tract infection*[Title/Abstract]). No limits were set on timeframe and languages.

Articles were manually retrieved. Three external reviewers (SL, LP, SG) screened each record and each report retrieved (title/abstract). Multiple reviewers (FS, MG, BE) worked independently at each stage of screening and an email process was used to resolve disagreements between screeners.

Data collection process was manually performed by three reviewers (SL, SG, LP) who independently worked. Synthesis methods included all the included studies which were qualitatively tabulated according to the study type (in vitro, in vivo, review) and main results relevant to describe the nature of the interaction between D-mannose and E. coli. Since the review does not regard clinical outcomes assessment, no risk of bias and no aggregated data could be estimated and synthesis of qualitative has been consulted [18,19].

2.3. Statements drafting

The research question has been discussed based on the Medical Device Coordination Group (MDCG) 2022-5 guidelines (Figure 2) [13] and art. 2 of the regulation 2017/745 (MDR) [20] 1 and available evidence on the mechanism of action of D-mannose for UTI prevention.

3. Results

3.1. Prisma results

A total of 33 articles were retrieved (Figure 2). Studies characteristics, results of individual studies and synthesis are reported in the supplement materials.

Figure 2. PRISMA flowchart of the selection process of articles that fulfilled the criteria.

Figure 2. PRISMA flowchart of the selection process of articles that fulfilled the criteria.

3.2. Delphi Panel results

After a kick-off web-based meeting, panellists were e-mailed a list of 17 statements. During Round I, experts were allowed to provide comments and suggest rephrasing or additional items. During Round 2, the statements that did not meet consensus after the first round were rephrased according to panellists’ comments: one statement was cancelled. During a final teleconference (29th November 2022), statements which had not reached a consensus after Round II were discussed and reviewed for the merit of inclusion and statements that had already reached a consensus were reviewed, discussed and validated: 3 statements were merged. In the end, 13 statements were drafted, 4 of which regarding definitions and the remaining specifically regarding alfa-D-mannose. Statements are reported in Table 1.

Box 2.  

Adhesion of pathogenic organisms to host tissues is the required step to initiate the majority of infectious diseases. It is mediated by lectins present on the surface of the infectious organism that binds to complementary carbohydrates on the surface of the host tissues. The bacterial lectins are typically in the form of elongated submicroscopic multi-subunit protein appendages, known as fimbriae (or pili).

In particular, the adherence of UTI to mannosylated protein called uroplakins (UPIa) on the bladder epithelium is the main mechanism of disease caused by uropathogenic E. coli (UPEC). This interaction occurs via the FimH adhesin, at the tip of the type I fimbria of E. coli, which is the virulence factor in UTI pathogenesis [28].

Adherence of P-fimbriated E. coli or of the isolated P fimbriae (also known as pyelonephritis-associated pili) to uroepithelial cells induces a two-way flow of biological crosstalk via the lectin bridge, affecting both partners. Following adherence, the target cells are activated, with resultant production of cytokines that produce acute inflammation and other symptoms of disease, while in the bacteria the interaction leads to up-regulation of signal transduction systems that allow responses to the changing environment. The FimH subunits of E. coli mediate not only bacterial adhesion, but also invasion of human bladder epithelial cells.

Over 90% of E. coli and other enteric bacteria express the type 1 fimbrial adhesin FimH, a lectin-like protein that binds specifically to terminal mannose or oligomannose residues on glycoproteins on a wide range of tissue. Without adhesion to mannosylated proteins on UPIa, UPEC would remain free in the urine and be removed from the bladder during urination, preventing the initial UPEC infection from progressing into a symptomatic UTI.

To bind to terminal mannose units UPEC produces multiple 3 μm-long rod-like structures on their surface known as type 1 pili, constituted by Fim A, FimF and FimG and indeed FimH (Figure below) [45].

FimH is comprised of two domains. The first is a C terminal pilin domain (FimHPD), which attaches FimH to the pilus rod through the neighbouring subunit FimG. The second FimH domain is the N-terminal lectin domain (FimHLD), which contains a mannose-binding pocket, that can bind to mannoside/oligomannoside sugars on the urothelial surface, mediating adhesion of UPEC to the urinary tract.

Figure 3. “Catch-bond” mechanism for the shear force-dependent binding of FimH to mannosylated urothelial surface. From Eris 2016 [29], permission requested.

Figure 3. “Catch-bond” mechanism for the shear force-dependent binding of FimH to mannosylated urothelial surface. From Eris 2016 [29], permission requested.

The two-domain structures of FimH allows the type-1 pilus to bind by a catch-bond mechanism. The catch-bond in FimH is biphasic: under moderate force (such as during urination) FimH binds to ligands with higher affinity. In fact, moderate mechanical flow force induces the separation of the FimHLD and FimHPL subunits, which switches the lectin domain from a low affinity state to a high affinity state with an up to 1000-fold higher affinity state

The relatively weak affinity of FimH under static conditions favours invasion of UPEC along the urinary tract while the high affinity of FimH under moderate flow conditions (urination) enables UPEC to be retained in the urinary tract during urination.

The panel concluded that, on the basis of the definitions, the mechanism of D-mannose does not involve a metabolic or immunological action while there is an uncertainty regarding the possibility of a pharmacological action.

As a matter of fact, D-mannose interacts with a cellular component present in the user’s body (the bacterial adhesin) and prevents a pathological process (i.e. the bacterial adhesion and infection). Besides, irrespectively of the fact that D-mannose activates or not an intracellular pathway in bacteria, the mechanism of action involves the prevention of the bacterial binding to uroepithelial cells and then the bacteria removal with the urines, with no interaction at all with the human tissue. From this prevention effect derives the well-established prevention clinical effect [53,54,55,56].

Uncertainty regards the features of the interaction between D-mannose and the bacterial adhesin and if they delineate a classical ligand-receptor interaction. On one side, the interaction is due to multiple bonds between specific atoms of D-mannose and specific aminoacids residues in a specific binding region of the adhesin. On the other side, such an interaction does not result in specific changes of the conformation of the bacterial protein and there is no activation of intracellular pathways important for the intended effect. From a regulatory perspective, the implications of this result, based initial assumptions, drives to maintain the possible classification of alpha-D-mannose as a medical device.

4. Discussion

From a regulatory perspective, the fact that the pharmacological action is uncertain (as for the panel votes), drives toward the classification of D-mannose as a medical device. An interaction between the product and the bacteria within the body occurs, it is reversible and dose-dependent, but its nature is inert because it does not induce a direct response activating or inhibiting body processes or restoring, correcting, or modifying the physiological functions in the human being. Moreover, it must be considered that the action of D-mannose takes place, even if inside the bladder, outside the epithelium on bacteria that have not yet invaded the urothelial tissue. Therefore, its mechanism of action is not directed to host structures but to structures (bacteria) external to the host's tissues.

Indeed, according to the EU jurisprudence (European Union Court (2009). Judgment C-140/07–hecht-pharma GmbH vs. staatliches gewerbeaufsichtsamt Luneburg [57]; European Union Court (2012). Judgment C-308/11–chemische fabrik kreussler and Co. GmbH v sunstar deutschland GmbH, formerly John O. Butler GmbH [58]) , products inducing a physiological effect cannot be automatically classified as medicinal products “by function” if the pharmacological, immunological, or metabolic effect is not demonstrated based on established scientific knowledge. Evidence does not demonstrate alpha-D-mannose pharmacological, immunological or metabolic action. Furthermore, the MEDDEV 2.1/3 rev.3 [59] and the updated version MDCG 2022-5 [13] define the pharmacological mode of action including two sequential steps: the interaction and the signal transduction pathway. The interaction by itself is not sufficient to determine the therapeutic effect [14].

The discussion around borderline products is quite active in the scientific community. The MDCG 2022-5 guidelines are the current instrument used to discuss the borderline nature of products. Using them in this research orientated the discussion toward a conservative hypothesis on the nature of the interaction between alpha-D-mannose and E. coli, being the guidelines the unique official document which introduces a specific definition of pharmacological effect. However, the guidelines do not have legal value and debate is ongoing in the scientific community around their completeness in the definitions provided. Nevertheless, regulatory considerations remain valid because guidelines include the MDR medical device definition.

Furthermore, Art. 2 related to medicinal products, included in the directive 2001/83/CE of the EU Parliament dated 6 November 2001- modified by the rule dated 31 March 2004, 2004/27/CE – does not apply to products whose functional quality of medicinal is not scientifically proven, although it cannot be excluded [21,60]. In other words, in presence of scientific evidence providing a complete demonstration of alfa-D-mannose as a medicinal product, it remains possible its classification as a medical device.

5. Conclusions

The regulatory framework is constantly changing. New regulations and legal judgments can add further consideration. A request pending at the European Court will answer to the specific case of a medicinal product with insufficient evidence of the drug and its regulatory classification. The judgment of the Court (Fourth Chamber) 3 October 2013, reports the fact that each Member State can have a different regulatory classification for the same product (medicinal product or medical device), demonstrating that there is not a unique opinion on classifications but, mostly, that it is not even required (ref Judgment of the Court (Fourth Chamber), 3 October 2013 Laboratoires Lyocentre v Lääkealan turvallisuus- ja kehittämiskeskus and Sosiaali- ja terveysalan lupa- ja valvontavirasto. Request for a preliminary ruling from the Korkein hallinto-oikeus [61]).

Finally, it is doubtful that further, new chemical evidence about the interaction a natural product like alpha-D-mannose and E. coli will be provided, due to the low affinity of the interaction: research is rapidly driving toward synthetic mannosides and their action in the prevention of UTI. The most promising explored class of mannose-based FimH inhibitors is alpha-d-mannopyranoside-based inhibitors [45]. In addition, multivalent FimH ligands have been developed to improve binding affinity and prodrugs to improve oral availability have been developed [62]. However, recent evidence demonstrated the previously unrecognized immunoregulatory function of d-mannose with potential application in immunopathology [6].