Molecular epidemiology and antimicrobial resistance patterns of Clostridioides difficile isolates in Algerian hospitals

Introduction : Clostridioides difficile is a major pathogen responsible for hospital-associated diarrhoea. This study investigated the molecular epidemiology and antibiotic resistance of C. difficile isolates in five Algerian hospitals. Methodology: Between 2016 and 2019, faecal specimens were collected from in-patients and were cultured for C. difficile . Isolates were characterised by toxin genes detection, Polymerase Chain Reaction (PCR)-ribotyping, Multilocus Sequence Typing (MLST), antimicrobial susceptibility testing against a panel of antibiotics, and screened for antimicrobial resistance genes. Results: Out of 300 patient stools tested, 18 (6%) were positive for C. difficile by culture, and were found to belong to 11 different ribotypes (RT) and 12 sequence types (ST): RT 085/ST39, FR 248/ST259, FR 111/ST48, RT 017/ST37, RT 014/ST2, RT 014/ST14, FR 247/new ST, RT 005/ST6, RT 029/ST16, RT 039/ST26, RT 056/ST34 and RT 446/ST58. MLST analysis assigned the isolates to two clades, 1 and 4. Clade 4 was more homogeneous, as it mainly included non-toxigenic isolates. Three toxin gene profiles were detected, two toxigenic, A+B+CDT-(33.3%) and A-B+CDT-(11%); and one non-toxigenic, A-B-CDT-(55.5%). All C. difficile isolates were susceptible to metronidazole, vancomycin and moxifloxacin. Conclusions: Overall prevalence of C. difficile in our healthcare settings was 6%. Antibiotic resistance rates ranged from 72.2% (clindamycin) to 16.6% (tetracycline). This study highlighted a relatively high genetic diversity in term of ribotypes, sequence types, toxin and antibiotic resistance patterns, in the C. difficile isolates. Further larger studies are needed to assess the true extent of C. difficile infections in Algeria.


Introduction
Clostridioides difficile, formerly known as Clostridium difficile, a Gram-positive, anaerobic, spore-forming, toxin producing bacteria, is the leading cause of healthcare-associated diarrhoea [1].The symptoms of C. difficile infections (CDI) can range from mild diarrhoea to pseudomembranous colitis or toxic megacolon, a severe form of the disease [2].The major risk factors for CDI are advanced age (≥ 65 years old), antibiotic exposure, a prolonged hospital stay, gastro-intestinal surgery as well as chronic conditions such as inflammatory bowel diseases [2].
The main virulence factors of C. difficile are the production of two major clostridial toxins: toxin A (TcdA) and toxin B (TcdB), encoded on a 19.6 kb chromosomally-located pathogenicity locus (PaLoc), which have cytotoxic and enterotoxic effects, respectively [3].However, certain strains of C. difficile produce a third toxin, called binary toxin (CDT), which acts as an actin-specific ADP-ribosyltransferase, encoded by the cdtA and cdtB genes, located outside the PaLoc [4].
The increase in the incidence of CDI reported worldwide over the last two decades was mainly attributed to the emergence of hypervirulent, multidrug-resistant strains, such as the epidemic NAP1/BI/027 strain [5].Compared to Europe and North America, CDI is a largely neglected disease in the developing countries and epidemiological data on C. difficile are scarce or lacking.In Africa, the highest prevalence of C. difficile was reported in Kenya (93.3%) [6] and the lowest in the Ivory-Coast (2%) [7], whereas in the Middle-East, the highest prevalence was recorded in Lebanon (82.9%) [8], and the lowest in Kuwait (0.5%) [9].In Algeria only one study was conducted reporting a prevalence of 6.9% [10].The wide variability in the above prevalence estimates is presumably due to a combination of factors such as study designs, type of population studied and C. difficile identification methods.
The aim of this study was to investigate the molecular epidemiology and antibiotic resistance of C. difficile in five hospitals located in three different provinces of Algeria.

Study design and sample collection
Unformed stool samples were collected from patients admitted to five hospitals in three different provinces, Batna, Ain Defla and Chlef; located in the East, Centre and West of Algeria, respectively, between January 2016 and January 2019.All patients included in the study developed diarrhoea, defined as 3 or more loose or liquid stools per day, or more frequently than is normal for the individual (as defined by the World Health Organization, http://www.who.int/topics/diarrhoea)[11], after three days of admission with or without previous antibiotic treatment.Infants under the age of 2 years old are excluded from this study due to the high asymptomatic carriage of C. difficile in this group [12].

Molecular identification and toxin genes detection
Genomic DNA was extracted using InstaGene Matrix Kit (Bio-Rad, Hercules, USA), following the manufacturer's instructions.C. difficile isolates were stored at -80 °C using Microbank mixed microbial storage vials (Pro-Lab diagnostics, Ontario, Canada).
For molecular characterisation of C. difficile isolates and toxin genes, a multiplex PCR assay was carried out according to the protocol of Barbut et al., 2019 [14], using seven pairs of primers targeting the following genes: tpi, (triose phosphate isomerase), tcdA (toxin A), tcdB (toxin B) cdtA and cdtB (binary toxin subunits), the PaLoc and tcdC (negative regulator for toxin expression) [15].C. difficile PCR-ribotype (RT) 027, was used as positive control.The amplicons were analysed using a high-resolution capillary electrophoresis detection system (HITACHI ABI 3500 Genetic Analyzer, Applied Biosystems, Massachusetts, USA).The results were visualised using GeneMapper Software version 5.0 (Applied Biosystems, Massachusetts, USA).

Polymerase Chain Reaction (PCR)-ribotyping
PCR-ribotyping was performed according to the protocol recommended by The European Centre for Disease Prevention and Control (ECDC), using primers designed by Bidet et al [16].For the amplification of the 16S-23S rRNA intergenic region, a capillary electrophoresis was performed using a Genetic Analyser (HITACHI ABI 3500, Applied Biosystems, Massachusetts, USA) and electrophoreograms were visualised using GeneMapper Software version 5.0 (Applied Biosystems, Massachusetts, USA).PCRribotype (RT) was determined using WEBRIBO database version 2.2 available at: https://webribo.ages.at/.When PCR-ribotyping profiles are unknown, the prefix "FR" was used (French reference laboratory internal nomenclature).

Statistical analyses
The data were coded using Excel 2013 (Microsoft, Redmond, WA, USA) and analyzed by R software (R Development Core Team, 2016).The distribution of prevalence of C. difficile between provinces, ages and sexes of patients was tested by Chi-Square test or Fisher's exact test.A level of p value < 0.05 was considered as statistically significant.

Detection of tcdA, tcdB, cdtA/B and tcdC genes
A PCR multiplex assay for the detection of tcdA, tcdB and cdtA/B genes, revealed the presence of three toxin genes profiles: six C. difficile isolates (33%) with intact tcdA and tcdB, deleted cdtA and the cdtB as a pseudogene (A+B+CDT-); two C. difficile isolates (11%) revealed a deletion in tcdA and intact tcdB, deleted cdtA and cdtB (A-B+CDT-); the remaining ten C. difficile isolates (55.5%) did not carry any of the toxin genes (A-B-CDT-) (Table 1).The same analysis revealed that the tcdC gene was present (without internal deletion) in all the toxigenic isolates and absent in all the non-toxigenic isolates.

Detection of antimicrobial susceptibility and antibioticresistance genes
Antibiotic-susceptibility data of the 18 C. difficile isolates are presented in Table 2.All isolates were susceptible to MTZ and VAN, the first line of antibiotics used for the treatment of CDI, and to MXF, a fourth-generation fluoroquinolone.Sequence analysis showed no substitution in the QRDR of gyrA of all our isolates.
A similar resistance rate against the macrolide antibiotics CLD was found (27.7%, 5/18) in the isolates of the ribotypes RT 29 (n = 1), RT 085 (n = 1) and FR 248 (n = 3); one of which (RT 29) was not found to carry the ermB gene.Two RT 085 (n = 1) and FR 247 (n = 1) isolates were resistant to ERY only.The latter isolate did not carry the ermB gene.Three isolates belonging to RT 014 (n = 1) and RT 017 (n = 2), which carried the ermB and tetM genes, were resistant to CLD, ERY and TET.
The remaining three isolates RT 014 (n = 1), RT 056 (n = 1) and RT 446 (n = 1) were susceptible to CLD, ERY and TET, and were not found to carry the ermB and the tetM genes.Interestingly, one isolate of the RT 039 harbored the tetM gene, but was susceptible to TET.
The two isolates of the RT 014, which belonged to two different STs, ST2 and ST14, also exhibited different antibiotic resistance phenotypes and genotypes; one isolate (RT 014/ST14) was resistant to CLD, ERY and TET, and carried the ermB and tetM genes; whereas the second (RT 014/ST2) was susceptible to these three antibiotics and did not carry the ermB and tetM genes.Of note, these two isolates were recovered from two different patients admitted to different wards of the same hospital.

Discussion
C. difficile has been identified as a leading nosocomial pathogen worldwide and the main causative agent of antibiotic-associated diarrhoea in humans [1].Although C. difficile infections were generally regarded as primarily healthcare-associated, and communityacquired, C. difficile infections have now emerged as a significant public health concern [2].Algeria is at the crossroads to Europe, Africa, and the Middle-East; and like in many developing countries, CDI is a largely neglected disease, and epidemiological data on C. difficile are scarce.There is, however, one previous study, based in two hospitals in one province [10].Therefore, this study was conducted to investigate the prevalence and the molecular epidemiology of C. difficile over a wider geographic region, in five study sites in three provinces of Algeria.Interestingly, the prevalence of C. difficile estimated here (6%) was similar to the previous Algerian study (6.9%) [10].
Compared to the previous Algerian study, which reported the detection of only four ribotypes [10], our study revealed a relatively larger diversity of PCR ribotypes.The most prevalent RT were RT 085 and FR 248 (n = 3, 16.7%, each), followed by RT 014/RT 017/FR 111 (n = 2, 11.1%, each) and RT 005/RT 029/RT 039/RT 056/RT 446/ FR 247 (n = 1, 5.6% each).All isolates of the same RT belonged to the same ST, with the exception of RT 014, which was shared by two different STs, ST2 and ST14 (1 isolate each), which is in agreement with a previous study [32].Among the above ribotypes, only RT 014 was previously reported in Algeria, but in a different hospital [10].
The ten C. difficile isolates were non toxigenic and belonged to ribotypes RT 039, RT 085, FR 111, FR 247 and FR 248; with RT 085 and FR 248 as the most prevalent (n = 3, 16.7%, each).The three isolates of RT 085 were detected in three different wards of the same hospital in the eastern province of Algeria (Batna), but were missing in the two other provinces; suggesting a possible distinct geographic distribution of this ribotype in Algeria.Although, the ribotype RT 085 was reported as more common in China [33], it was rarely reported in other countries.
It is also worth mentioning that all the three isolates belonging to the unclassified ribotype FR 248 were recovered from children, that were admitted to two different hospitals from two provinces (Ain Defla and Chlef), and as such, the possibility of an association of this ribotype with children is plausible.
The remaining non toxigenic ribotype isolated with a lesser frequency (n = 1), RT 039, was previously reported as most common in patients with cystic fibrosis in Western Australia [34], and was also detected in health care settings in Iran [35], and Kuwait [36].Surprisingly, several isolates of RT 039 from Iran were found to carry the toxin genes [35].In addition, isolates of RT 039 were also recovered from animal samples in the Netherlands [37] and Egypt [38].
Toxigenic isolates accounted for 44.4% (8/18) of the total and were shared between six ribotypes, 005, 014, 017, 029, 056 and 446; among which, those belonging to RTs 014 and 017 were the most frequent (n = 2, 11.1%, each).Isolates of the RT 014 were the most prevalent ribotype in many European countries, where it was reported as responsible for CDI outbreaks in humans, and also commonly associated with animals and different environments [37,[39][40][41][42].The RT 014 was also detected in several countries in the Middles-East, Iran [35], Lebanon [8] and Qatar [28]; whereas in the African continent it was reported only in Algeria [10] and South Africa [43].
The other most prevalent ribotype in this study, RT017, which is characterized by a deletion in the tcdA gene and the absence of binary toxin genes and, therefore, A-B+CDT- [44], is the predominant ribotype in Asia, and has also caused major outbreaks of CDIs in several countries around the world [45,46].To date, the only African country that reported the ribotype RT 017 is South Africa [43,47], but it has not been detected so far in any of the Middle-eastern countries.
The toxigenic PCR-ribotype RT 029 was previously reported as one of the most frequent RTs among hospitalized patients in Iran [35], and was also isolated from humans in Egypt [48] as well as from humans and animals in Costa Rica [49].
The toxigenic isolate belonging to the PCRribotype RT 056 detected in this study, was commonly isolated from humans, cattle, vegetables and the environment in Australia [42].Prior studies reported that RT 056 was frequently associated with complicated CDI in hospitalized patients in Europe [39,50].The only report to date of this PCR-ribotype in Africa comes from Zimbabwe [25], whereas in the Middle-East it was reported in Qatar [28] and Kuwait [9,36].
Although the toxigenic RT 005 identified in this study is among the most common ribotypes in Europe [51], it was isolated with a low frequency in a study from Ghana [52], and has not been documented so far in the Middle-East.
Importantly, both this and the previous Algerian study failed to detect the hypervirulent ribotypes RT 027 or RT 078.It must be noted that the ribotype RT 027 was not reported in the African continent, whereas, in the Middle-East, it was detected, albeit with low frequency, in Iran (n = 14) [26,53,54], Saudi Arabia (n = 4) [55] and Qatar (n = 1) [28].Similarly, there is very little data on the ribotype RT 078 in Africa and the Middle-East, except for two reports from Egypt (n = 6) and Kuwait (n = 9) [36,48].
It is worth mentioning that the detection of the toxigenic ribotypes RT014, RT017, RT029 and RT 056 in this study is important and interesting from an epidemiological point of view, given that these ribotypes were reported to be either responsible for CDIs in several countries around the world (RT017, RT 014 and RT 056) [50], or commonly associated with animals (RT 014, RT 029 and RT 056) [56], raising concerns about their potential zoonotic transmission.
Antibiotic susceptibility test results have shown that all 18 C. difficile isolates were susceptible to VAN and MTZ, the treatment of choice for CDI [57], as well as to MXF, a fourth generation fluoroquinolone [58], which is in line with the results of several other studies [59].
Thirteen MLSB-resistant isolates (72.2%, 13/18 ), carried the ermB gene, and were resistant to CLD and/or ERY; whereas two (11.1%,2/18 ) MLSB-resistant isolates were ermB-negative but resistant to either CLD or ERY; suggesting that the MLSB resistance in these isolates might be conferred by other mechanisms; which is in agreement with previous studies [20,60].
We noticed that resistance to TET was always associated with co-resistance to CLD and ERY, as is the case for three isolates belonging to RT 014 (n = 1) and RT 017 (n = 2), which harboured both the tetM and the ermB gene.The other TET resistance genes investigated in this study (tetO, tetB[P], tet0/32/0, tet40, tetA[P]) were not detected in any of our 18 isolates.Surprisingly, one isolate, member of RT 039, was susceptible to TET despite carrying the tetM gene.Given that tetM is the predominant TET resistance genetic determinant in C. difficile, the exact mechanism behind this peculiar phenotype is unclear at this stage; it is possible, however, that the tetM gene in this isolate was inactive due to a mutation.
The high rate of resistance of RT 017 to many antimicrobial agents has been largely documented in several studies, and considered as a major contributing factor to the success and dissemination of this ribotype throughout the world [61,62].
There are a number of limitations within this study that needs to be highlighted; first, and most important, our study lacked clinical patient data; second, this is a study based on a small sample size of isolates; third, C. difficile isolates were collected in three geographical areas and five hospitals, which may limit the generalization of the C. difficile prevalence estimates to the whole country; fourth, lack of antibiotic susceptibility testing against other important antibiotics used for the treatment of CDI such as fidaxomicin and rifaximin.Clearly, larger studies, over wider geographical area and larger number of study sites are merited.

Conclusions
The present study revealed a moderate prevalence of CDI (6%), with a relatively high diversity of C. difficile isolates, some of which were toxigenic.All isolates were susceptible to VAN and MTZ; whereas a high proportion of the isolates showed resistance to CLD and/or ERY.Although well-known hypervirulent C. difficile strains such as RT 027 and RT 078 were not detected in this study, our findings highlight the significance of this pathogen in a sample of the Algerian population, and therefore, an active surveillance of CDI is crucial in order to have a more generalized estimation of the burden of this disease in the country.

Table 1 .
Molecular characterisation of C. difficile isolates in the study.
).The three unrecognized isolates, FR 111, FR 247 and FR 248, detected in this study corresponded to ribotypes maintained in the internal database of the French National Reference Laboratory for C. difficile.Our C. difficile isolates were classified into two MLST clades, 1 and 4; (Table1) clade 1 was more heterogeneous and consisted of a diverse set of isolates, RT 005/ST6, RT 14/ST2, RT 14/ST14, RT 029/ST16,

Table 2 .
Antimicrobial resistance patterns of C. difficile isolates.