The incoming Medical Devices Regulation (MDR) No 2017/745 (Regulation (EU) 2017/745), fully in force since 26 May 2021, has introduced several innovative aspects throughout the life-cycle of a medical device (MD), especially for high-risk MDs, including: 1) mandatory clinical data (introducing, albeit not defining, the concept of sufficient clinical evidence); 2) proactive post-marketing clinical follow-up (PMCF) plan, comprising a Periodic Safety Update Report (PSUR) to be updated at least annually; 3) a new international Unique Device Identification (UDI) system and a publicly accessible European database (EUDAMED) to increase traceability and transparency of MDs (Antich-Isern et al., 2021). This European legal framework mirrors the regulatory scenario of medicinal products, albeit with notable differences (Van Norman, 2016a; Van Norman, 2016b; Van Norman, 2018; Cipriani et al., 2020; Naci et al., 2020) (Table 1).

While promoting a real proactive pre- and post-marketing risk-benefit assessment, the MDR poses a number of challenges for stakeholders, including manufactures, notified bodies and researchers (pharmacologists and toxicologists), and formally recognizes the role of “medical devices made of substances” (MDMS), especially addressed by Rule 14 and 21. The former covers MDs incorporating, as an integral part, a substance which, if used separately, can be considered to be a medicinal product, albeit with an ancillary action (e.g., drug-eluting stents); the latter identifies substance-based MDs introduced via a body orifice such as the gastrointestinal tract or applied to the skin.

Two key issues should be tackled for MDMS. First, the non-targeted mechanism of action “not pharmacological, immunological or metabolic” of the MD is different from the “therapeutic effect” and should be determined (from the literature or experimentally) (Racchi et al., 2016; Racchi and Govoni, 2020), as exemplified by the case of lubiprostone and glycerine as laxatives (Sardi et al., 2018). Second, pharmacokinetic and toxicological data, including biocompatibility, deserve careful assessment, by demonstrating local and systemic exposure to the (ancillary) medicinal substance, its potential impact on safety, interactions with other substances, and relevant disposition of metabolites, taking into account individual variability.

The most challenging aspect lies on MDs made of complex natural products, encompassing food supplements, botanicals, and herbal medicinal products, which do not fit the so-called key-lock model due to their complex multitarget mechanism (Bilia et al., 2021), with notable impact on environmental sustainability. “System biology” has been regarded as a new scientific paradigm to investigate these multiple interactions between complex substances and the human body.

The regulatory definitions of pharmacological, immunological, or metabolic means given in Meddev 2.1/3 rev 3 (European Commission, 2009) have been recently amended in the Guidance on borderline between medical devices and medicinal products (MDCG 2022–5) with the aim at adding more precision to such definitions (Supplementary Table S1). Notably, the term “receptor”, historically associated with the pharmacological mean, is no more the focus and the “pharmacological mean” was acknowledged to encompass several molecular interactions (e.g., covalent bonding, H-bonds, electrostatic, and van der Waals forces) with constituents of the human body. It is likely that this new definition will further fuel the discussion but, an “interaction at the molecular level”, is still understood as a specific interaction that should be demonstrated and identified unequivocally. As proposed by Racchi et al. (2016) and by Bilia et al. (2021), when a specific primary target cannot define the mechanism of action, this should not be regarded as a pharmacologic mean. Moreover, the interaction itself is not sufficient to determine the therapeutic effect but should trigger a subsequent signal transduction pathway resulting in “initiation, enhancement, reduction or blockade of physiological functions or pathological processes” (Leone, 2022). Note 5 also reads that “Although not an exhaustive criterion, the presence of a dose-response correlation is indicative of a pharmacological, metabolic or immunological mode of action”.

Moving from single active substances to very complex mixtures, such as the case of natural products, the correct interpretation of “pharmacological mechanism” is even much more challenging since it is very hard to specifically identify, describe and measure all the interactions each constituent can have each other and within the body.

Typically, the claimed therapeutic effect of MDs is achieved by a physicochemical mechanism of action. However, based on Rule 21, the MDR expressly recognizes the existence of MDMS acting systemically other than locally, de facto acknowledging that the concept of non-pharmacological mechanisms of action may go beyond chemico-physical means. Even in such cases, the effect should not be achieved through specific ligand-primary target interactions eliciting definite cell signaling changes.

Rule 21 represents the main novelty of the MDR paving the way for a strong therapeutic innovation while establishing a robust, regulatory framework as guarantee of the efficacy and safety of these products. Notably, MDMS can no longer be classified as class I products and will face a conformity assessment by the Notified Body; in addition, for systemically absorbed products, the scientific opinion of a medicinal competent authority or of the European Medicines Agency (EMA) will be asked and a case-by-case evaluation on whether these product fall within the MDR will be provided.

However, the practical experimental approaches for classifying MDMS under the definition of “non-pharmacological mean” remains a major challenge.

Historically, ethnopharmacological research on complex natural products has been aimed at identifying and isolating the bioactive principle to be further developed as a medicinal product once the target and pharmacological mechanism of action was identified. A typical starting point is the measurement of a given biological activity using in vitro phenotypic assays followed by target deconvolution (Terstappen et al., 2007). This approach is particularly valuable for natural products since bioactive molecules can be identified without preconceived molecular mechanisms (Chang and Kwon, 2016).

In the case of complex natural substances, investigations usually begin with a crude extract tested in vitro assays followed by biological activity-guided fractionation until a single active principle or structurally related compounds, are isolated (Najmi et al., 2022). The mode of action of the main active ingredient is then characterized by pharmacological and molecular tools (receptor binding studies, downstream pathways activation/inhibition or enzymatic activity modulation).

Several important guidelines for performing rigorous pharmacological research into natural products were recently published (Izzo et al., 2020): among those, the selection of appropriate concentrations or doses reflecting real-life exposure patterns and the evaluation of the mechanism of action at such thresholds, are particularly valuable also for MDMS research. Moreover, since natural products are often very complex mixtures of different chemical entities that can be prone to geographical, seasonal, and environmental variations, the determination of the chemical composition and its standardization, is another crucial aspect to be addressed (Mattoli et al., 2022). Lack of phytochemical qualitative and quantitative analysis of a particular natural product, may affect the potency, may result in unreproducible effects, whether they are achieved through a pharmacological or a non-pharmacological mechanism of action, and ultimately, may impact its placement in the market and its clinical applications.

When the biological activity of a multicomponent mixture is investigated, the phenotypic change assayed should be observed in the presence of the combinatorial components acting in concert in multiple, but not individually identifiable targets, otherwise, the active principle(s) should be identified, and the mechanism of action further characterized. Synergistic effects can be demonstrated against the phenotypic changes observed with single ingredients: Long et al. (2015) elucidated the predominant combinatory compounds of an herbal extract contributing to an anti-inflammatory effect and demonstrated additive effects by comparing dose-effect curves for single compounds alone and for multiple combinations. However, this strategy assumes that the underling mechanism of action is pharmacological and its relevance and applicability for MDMS is uncertain.

Within system biology, network pharmacology (NP) has emerged as a paradigm shift for tackling drug attrition during drug discovery and pre-clinical development (Hopkins, 2008; Berg, 2014; Gomez-Verjuan et al., 2019). By accounting for and understanding multiple drug-target-disease interactions, multi-component and multi-target agents represent a more effective and less toxic therapeutic approach in complex diseases, as compared to single-target therapy (Hopkins, 2008; Medina-Franco et al., 2013).

In the last decade, NP was especially exploited to elucidate potential targets and toxicity of natural products, including traditional herbal medicines, and detect possible synergism of their constituents (Li et al., 2011; Yuan et al., 2017; Lee et al., 2019). Thus, can NP help deal with the challenge of benefit-risk assessment for SBMD of natural products imposed by rule 21?

With regards to clinical benefit, the non-targeted mechanism of action of MD imposes a careful reflection on the use of NP, which, by definition, usually exploits druggable targets. A network should be designed based on physicochemical and/or physical mechanisms of action that cannot be pinpointed at the single target/receptor level, as exemplified by MDMS used for gastrointestinal disorders (Bilia et al., 2021).

With regards to safety, both medicinal products and MDMS need a full characterization of their potential toxicity. MDMS made by complex mixtures of substances make harder the definition of the toxicological profile, which cannot be considered the sum of the toxicity of single components (Racchi and Govoni, 2020). In this intricate scenario, where the target is considered the function instead of the receptor, the global safety evaluation could be predicted through a network toxicological approach, as applied for traditional Chinese medicine (Zhang et al., 2019). The network approach, also applied to system biology, could take advantage from the so-called “omics-based” technologies (Thomford et al., 2018) to booster innovation and development in the field of MDMS made of natural products.

Several chronic, multifactorial disorders with unsatisfying medical treatment can benefit from MDMS regulation; this is the case of treatment of some gastrointestinal illness (Corazziari, 2020; Bilia et al., 2021) or metabolic disorders (Guarino et al., 2021), even in the pediatric population (Strisciuglio et al., 2021; Stagi, 2022). In this scenario, MDR provides opportunities for therapeutic innovation with complex substances while guaranteeing patients with standards of efficacy and safety. Another therapeutic opportunity for complex substances, either natural or synthetic, is Rule 21, and its first and second indent referring to MD which need to be absorbed to achieve their intended purpose.

However, the practical approaches and scientific evidence that is required to demonstrate a non-pharmacological mechanism of action are still blurred: current paradigms, including system biology and network pharmacology are applied to the study of the complexity of the biological activities elicited by both drugs and mixtures of natural substances and how they can be applied to the study of the mechanism of action of MD is still uncertain. The difficulties in delineating the boundaries between pharmacological and non-pharmacological mechanism of action and the regulatory classification of even a single substance, such as citicoline in glaucoma (Marchesi et al., 2022), exponentially increase in the case of complex mixtures of substances.

We encourage pharmacologists and toxicologists to take advantage of this journal section as a forum for open discussion of novel proposals on experimental set ups, for reviewing current research and sharing opinions and experiences on both preclinical and clinical research in this area.