**1. Introduction**

Over the last hundred years, biochemical discoveries have made it increasingly possible to characterize the metabolic pathways in our bodies, develop new drugs, and monitor human nutrition and lifestyle. Though our knowledge is increasingly broad, it remains divided into specific areas, such as the characterization of genetic make-up or transcriptional factors underlying the expression of essential proteins involved in specific physiological or pathophysiological processes. In this context, metabolomics has come into its own to mend the cracks between the different disciplines hitherto used to study our biochemical mechanisms [1]. Considering that a single change in a DNA base can lead to the observation of alterations in metabolite concentrations of up to 10,000-fold changes [2], metabolomics represents a highly sensitive probe for depicting our phenotype. Helped by the development of new analytical technologies for obtaining and processing biochemistry data, metabolomics as an omics discipline is under constant development. In the last 20 years alone, more than 5000 papers have been published on the subject, making it one of the fastest-growing disciplines [3].

**Citation:** Petrella, G.; Montesano, C.; Lentini, S.; Ciufolini, G.; Vanni, D.; Speziale, R.; Salonia, A.; Montorsi, F.; Summa, V.; Vago, R.; et al. Personalized Metabolic Profile by Synergic Use of NMR and HRMS. *Molecules* **2021**, *26*, 4167. https:// doi.org/10.3390/molecules26144167

Academic Editor: Robert Brinson

Received: 14 June 2021 Accepted: 7 July 2021 Published: 8 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

The most used analytical platforms in metabolomics are chromatography-mass spectrometry (LC–MS, GC–MS, CE–MS, and IMS–MS) and NMR spectroscopy. As reported in many papers, these two methodologies have several features that make them complementary [4]. For example, MS techniques are highly sensitive and allow for the detection of thousands of features at different concentration ranges, potentially expanding the description of a metabolic profile in detail with just a few microliters of sample. However, the identification of compounds by MS is a more complex process than by NMR. Indeed, the metabolite identity is solved by measuring the mass-to-charge ratio (*m*/*z*) of the ionized molecule and/or its ionized molecular fragments and then comparing them with reference spectra and/or using analytical standards [5]. Furthermore, not all MS techniques have the same degree of reproducibility; this is mainly the case for LC–MS measurements, which yield less reliable metabolite quantifications [6].

Unlike mass spectrometry, NMR is not a destructive technique and, in many cases, requires minimal sample preparation. The ability to determine the identity of a compound with a single analysis (1H-NMR) can be very accurate and fast for concentrated compounds or those that give signals in non-crowded regions of the spectrum. However, it has a lower sensitivity than MS [3], making it possible to only quantify a portion of the metabolome.

Thus far, few studies have described a combined use of both techniques, and these have directed their attention towards the development of statistical methods for weighing the two datasets [4] or for the structure determination of new compounds in commonly studied biofluids [7]. However, given the high complementarity of the two techniques, it should be beneficial to combine the data separately obtained with NMR and MS to improve the ability to classify and quantify the "metabotypes" under investigation [8].

One of the main limitations of using NMR is the relatively low number of accurately quantifiable metabolites in particularly complex mixtures like urine. For example, a study performed at two different fields (600 and 700 MHz) starting from a list of 151 metabolites that are potentially quantifiable in urine showed that only 50 presented data strongly correlated between the values obtained at the two magnetic fields [9]. This result represents a limit of quantifiable compounds in urine using NMR, and most studies in the literature have used a dataset of this size [10–15]. However, in one case, it was possible to reach 209 quantified metabolites [16], but only a fraction was detected in more than 80% of the samples.

Simultaneously identifying a metabolite by both NMR and MS would maximize the advantages for biomarker discovery by increasing the number of quantified metabolites in all samples and the accuracy of the measured concentrations. In our opinion, the method that has best combined MS data with NMR is the one developed by Nicholson et al. [17]. The authors named this strategy Statistical Heterospectroscopy (SHY) and showed that it is possible to correlate chemical shift and *m*/*z* data when a cohort of samples is considered. This concept revealed a new perspective to cross-reference NMR and MS data and to get the best of both techniques. However, the correlation was attempted with regions of the NMR spectrum, limiting the number of identifiable metabolites and obtaining only relative levels instead of concentrations. Our idea is to use the SHY concept to develop a novel strategy of the MS-assisted deconvolution of NMR spectra to extend the number of urinary metabolites quantified in their absolute rather than relative levels. We show how the synergistic use of both analytical methodologies can help to achieve this goal, taking the determination of metabolite concentrations in human urine as a specific case. We call this approach: *SY*nergic use of NMR and HRMS for *MET*abolomics (SYNHMET). Using SYNHMET, it was possible to obtain a complete dataset comprising 165 urinary metabolite concentrations for nine controls, six patients affected by chronic cystitis, and thirty-one bladder cancer patients.

#### **2. Results**

#### *2.1. Metabolite Levels in Urine Acquired by NMR and HRMS*

The SYNHMET method was first applied to quantify metabolites in urine samples from 46 subjects, divided into three groups: nine healthy controls (CTRL), six patients with

chronic cystitis (CC), and thirty-one bladder cancer patients (BC). The 1H-NMR dataset was acquired using a 600 MHz Bruker Avance spectrometer (Bruker, MA, USA). The HRMS dataset was acquired using a UHPLC–high-resolution mass spectrometry (UHPLC– HRMS) analysis system coupled to an Orbitrap QExactive™ mass spectrometer (Thermo Scientific™, MA, USA) equipped with a HESI source operated in the positive and negative ion modes. In addition, we used two different chromatographic conditions: reverse-phase (RP), which allows for the separation of metabolites based on hydrophobic interactions, and hydrophilic interaction liquid chromatography (HILIC), enabling the analysis of polar compounds. Combining two ion modes and two chromatographic conditions allowed for broad coverage of urinary metabolites, an essential feature in an untargeted approach. After the MS analysis, 10,497 hits were obtained. Information about a matched formula, exact mass, retention time, and relative intensity was available for each hit, and in many cases, so was a putative name (Table 1).

**Table 1.** Examples of two hits obtained with UHPLC–HRMS and the information available for each of them.


<sup>1</sup> HC+: HILIC positive; RP-: reverse phase negative.

## *2.2. Extraction of Urinary Metabolite Concentrations by SYNHMET*

The proposed workflow for the SYHNMET method applied to the human urine samples described above is shown in Scheme 1. The following chapters illustrate its application focused on a specific region of the 1H-NMR spectrum.

**Scheme 1.** Workflow of the SYNHMET method that allowed for the identification and accurate quantification of a significant number of metabolites in 46 urine samples in this application.
