Role of Institut Hospitalo-Universitaire Méditerranée Infection in the surveillance of resistance to antibiotics and training of students in the Mediterranean basin and in African countries

Surveillance of antibiotic resistance has become a public global concern after the rapid worldwide dissemination of several antibiotic resistance genes. Here we report the role of the Institut Hospitalo-Universitaire Méditerranée Infection created in 2011 in the identification and description of multidrug-resistant bacteria thanks to collaborations and training of students from the Mediterranean basin and from African countries. Since the creation of the institute, 95 students and researchers have come from 19 different countries from these areas to characterize 6359 bacterial isolates from 7280 samples from humans (64%), animals (28%) and the environment (8%). Most bacterial isolates studied were Gram-negative bacteria (n = 5588; 87.9%), mostly from Algeria (n = 4190), Lebanon (n = 946), Greece (n = 610), Saudi Arabia (n = 299) and Senegal (n = 278). Antibiotic resistance was diversified with the detection and characterization of extended-spectrum β-lactamases, carbapenemases and resistance to colistin, vancomycin and methicillin. All those studies led to 97 indexed international scientific papers. Over the last 6 years, our institute has created a huge network of collaborations by training students that plays a major role in the surveillance of resistance to antibiotics in these countries.


Introduction
During the last decade, antibiotic resistance has become one of the major public health priorities in the world [1] because of the emergence of new mechanisms of resistance. Moreover, the massive media coverage has tended to predict of thousands of human deaths every year [2]. However, recent epidemiologic data from our institution demonstrate that the level of antibiotic resistance for the most common bacterial species of clinical interest did not significantly change over the last 15 years in Marseille, France [3,4]. Similarly, we found a huge disparity between mortality attributable to antibiotic resistance using simple model estimations and empirical data of true deaths in our institution [5]. Data on the level of antibiotic resistance in Europe show disparities between countries and bacterial species for certain antibiotics; for example, resistance to carbapenems is much more frequent in Romania, Italy and Greece [3]. It appears from those studies that a better understanding and surveillance of antibiotic resistance at the local and national levels is critical to manage antibiotic-resistant bacterial infections in the future [5]. However, data on antibiotic resistance and surveillance of the emergence and spread of new mechanisms of resistance in the Mediterranean basin and in African countries were lacking in most of those countries until now.
Here we report the specific and unique role of the Institut Hospitalo-Universitaire Méditerranée Infection (IHU-MI), created in November 2011, in the identification, description and surveillance of multidrug-resistant bacteria thanks to collaborations among and training of students coming from the Mediterranean basin and from African countries in our institute. The majority of students who come to our institute for surveillance and analysis of antibiotic resistance came with their own bacterial isolates from their countries.

Methods
This study analyses data collected from 2011 (the date of creation of the IHU-MI) through the completion of this article in February 2018. The number of students by level of graduation and by country of origin per year was sorted from our administrative database of students and scientific visitors during the study period from the team dedicated to antibiotic resistance research (JMR team). Students from the Mediterranean basin and Africa were counted from this primary list and were sorted by level of graduation (master's degree, PhD, postdoc and scientific visitors) and by country. Because some students stayed at our institute both for master's and PhD courses, we deduplicated the total count. The number of students present per year in the team was also calculated from this list to show the student kinetics of reception per year.
All students facing antibiotic resistance in their country came with their own isolates to analyse them as a course training. Most of them continued to collaborate with our institute, resulting in real-time surveillance of antibiotic resistance according to their field of research (humans, animals or the environment). Initially there was no rationale for the recruitment and analysis of the samples because no data existed at the beginning of this network. Now, however, the follow-up of antibiotic resistance is mainly focused on the current antibiotic resistance situation. For each epidemiologic study, the number and type of samples and/or bacterial isolates and the country of origin were counted, and data were presented in a single table, with all data provided by country.
Antibiotic resistance for each sample or bacterial isolates was studied using the same procedure. Antibiotic resistance was assessed either directly from samples by PCR or from bacterial isolates by culture and molecular assays. The first step consisted of sample culture and isolation of strains on specific agar media: Columbia agar with 5% sheep's blood, trypticase soy agar or MacConkey (bioMérieux, Marcy l'Etoile, France) with or without addition of antibiotics. All collected strains were subjected to matrix-assisted laser desorption/ ionization time-of-flight mass spectrometry (MALDI-TOF MS) for identification [6]. Antibiotic susceptibility testing (AST) was performed using the disc diffusion method on Mueller-Hinton medium agar for phenotypic characterization of the mechanism of resistance. Specific panels of antibiotics were tested according to the bacteria species (e.g. Enterobacteriaceae, nonfermentative Gram-negative bacteria, Gram-positive bacteria). Then AST results were interpreted according to European Committee on Antimicrobial Susceptibility Testing guidelines [7]. Genotypic identification of resistance genes were screened by real-time quantitative PCR and confirmed by standard PCR and sequencing when necessary, and sequences were analysed using ARG-ANNOT software [8] to identify the specific antibiotic resistance gene. Multilocus sequence typing was performed to evaluate genetic relatedness of strains. If necessary, a whole genome sequence study was performed to obtain the complete resistome of a strain [9] or to describe the genetic environment of an antibiotic resistance gene [10,11].
Each student hosted by our institution received specific training for the study of antibiotic resistance (MALDI-TOF MS, AST, molecular training, genomics, bioinformatics) and presented the progress of their work and their results every week so that we could prepare tables and figures to be used for publication. Finally, each student was trained by the senior member of the team (JMR) to write their scientific papers and to create their own bibliography on the topic. Most of them also wrote a review on their topic while writing their PhD thesis. Weekly seminars or bibliographic sessions were also provided each Friday to improve students' knowledge in the field. The number of indexed international scientific papers per type of sample and per country was also calculated on the basis of published and submitted papers on antibiotic resistance during the study period.

Results
Since the creation of this institute, the JMR team has welcomed a total of 126 students or visiting scientists, including 95 deduplicated students (75.4%) from academic exchanges with 19 countries from the Mediterranean basin, Africa and Middle East. The number of students present in a given year has significantly increased during the study period (ten students in 2011, 15  Most of the students are from Algeria (52, 55.3%), followed by Lebanon (12, 12.8%), Senegal (7, 7.4%) and Tunisia (5, 5.3%). All these 95 students were from Europe (Spain, 3 students, 3  . Students trained at the IHU-MI will return to their country of origin and continue to work in this field with our institute, which is now identified as the core laboratory for surveillance of antibiotic resistance and further analysis of new bacterial isolates from those countries. A total of 7280 samples from human (n = 4657; 64%), animal (n = 2058; 28%) or environment (n = 565; 8%) from 15 different countries were analysed during the study period ( Fig. 1(A)). More than half of those samples came from Algeria (n = 4190; 57.6%), followed by Lebanon (n = 946; 12.9%), Greece (n = 610; 8.4%), Saudi Arabia (n = 299; 4.1%) and

Continued
All these studies allowed the detection and characterization of specific antibiotic resistance genes in multidrug-resistant bacteria from these genera and led to 97 scientific international indexed publications. Table 1 lists the publications and findings of specific antibiotic resistance determinants by country and bacterial species.
The main antibiotic resistance determinants detected and characterized were extended-spectrum β-lactamases (ESBLs) and carbapenemases including bla CTX-M , bla SHV , bla TEM [12], bla OXA-48 [13], bla NDM [82] and bla VIM [14] genes. Genes encoding for resistance to aminoglycosides were also reported, including, for example, armA or aac(6 0 )-Ib in Acinetobacter baumannii [81]. Resistance to colistin mediated by the newly plasmid-mediated mcr-1 gene in human and animal isolates has been tested to date in 21 studies from eight countries (Algeria, Greece, Israel, Lebanon, Nigeria, Saudi Arabia, Senegal and Spain), leading to ten scientific publications (Table 1). An overview of the global distribution of the main findings of antibiotic resistance determinants in the 7280 samples studied per country and type of samples is provided in the map in Fig. 2.

Discussion
Here we show the unique role of IHU-MI in training about 100 students working in the field of antibiotic resistance from the Mediterranean basin and Africa over the last 6 years. This has led to the description and surveillance of new mechanisms of resistance to antibiotics in 15 various countries reported in 97 scientific publications, including 24 different peer-reviewed journals. The majority of the publications have reported the first detection of antibiotic resistance genes, mainly ESBLs [47], carbapenemases [15,16,31] and the mcr-1 plasmid-mediated colistin resistance gene [17,18,22] in these countries in both humans and animals. One of the main contributions in the field is the description of a strong link between antibiotic consumption in animals and emergence and spread of antibiotic resistance genes in animals as well as the transfer to humans [83].

S60
New Microbes and New Infections, Volume 26 Number S1, November 2018 Antibiotics are widely used in agricultural settings in these countries, without clear control policies; this situation has affected human health and is implicated in the evolution of new mechanisms of resistance [84]. Epidemiologic descriptions are essential, and our results confirmed that surveillance should continue in Africa and in the Mediterranean basin to monitor and control the emergence and spread of antibiotic resistance genes. Thanks to these students and their training at the IHU-MI, the institute has created a unique collaborative network for surveillance and study of antibiotic resistance in Africa and in the Mediterranean basin because most of these students returned to their country of origin and created microbiology laboratories to study and survey antibiotic resistance in collaboration with the IHU-MI institute. Further engagements with key individuals are ongoing to create new partnerships to study antibiotic resistance in humans and animals in Italy, Morocco, Turkey and in the Balkans, and should be reinforced in Egypt and Libya as well as the Middle East, although the political situation is currently complex. The current migrant crisis in Europe should prompt us to survey antibiotic resistance in humans and animals from these countries, including Syria and Iraq, to avoid the possible spreading of specific clones, as previously reported in Greece [85,86] and Israel [87] for Klebsiella pneumoniae carbapenemase producers.
Because of its special location as a seaport in the Mediterranean basin, Marseille has historically always been a critical place for the entrance of infectious diseases such as plague or cholera [88]. Because antibiotic-resistant bacteria and antibiotic resistance genes that could spread in the Mediterranean basin do not have borders, the IHU-MI in Marseille plays a critical role in the surveillance of resistance in these areas as well as in African countries that historically have links to France. Thus, over the last 6 years, the institute has become a reference centre for the surveillance of antibiotic resistance and the training of students from countries in the Mediterranean basin and Africa. Such a collaborative network will expand in the future, permitting real-time surveillance of antibiotic resistance determinants that may emerge and spread in these areas [89].