Prevalence and Genetic Diversity of Clostridium perfringens Isolates in Hospitalized Diarrheal Patients from Central China

Introduction

Clostridium perfringens (C. perfringens) is an important pathogen, causing human gastrointestinal (GI) diseases, particularly in the hospital environment. Recently, subtypes of C. perfringens have been classified into seven toxin types (A through G) based on their ability to produce the major lethal toxins, alpha, beta, epsilon, iota, enterotoxin (CPE) and NetB.¹ CPE encoded by cpe gene, is a 35-kDa polypeptide that binds to claudin receptors on enterocytes to form pores, disrupting the intercellular claudin tight junctions and causing intestinal disease symptoms.² CPE-positive C. perfringens subtypes are responsible for causing food poisoning (FP) and non-foodborne gastrointestinal (GI) diseases such as AAD and SD.^3–5 Type F represents the formerly called CPE-positive isolates of C. perfringens type A and produces CPA and CPE toxins. Recently, Azimirad et al reported that 13.3% of patients with AAD carried type F isolates, demonstrating the importance of type F in the development of AAD.⁴ Additionally, type F was reported as an essential infection source for asymptomatic carriers and foodborne diseases.⁶ Clostridium perfringens type C and type D also express CPE,¹ and type C can cause enteric diseases characterized by vomiting, diarrhea, and abdominal cramps. Type C can also cause fatal intestinal necrosis in humans and other animals.^7,8 Apart from CPE, beta2 toxin encoded by the cpb2 gene has been associated with enteric disease in humans and necrotic enteritis in chickens and many other animal species.^7,9,10 A recent study suggested that beta2 toxin was significantly associated with children suffering from autism spectrum disorders (ASD).¹¹

Multi-locus sequence typing (MLST) is a commonly used technique for the typing of human, animal and foodborne pathogens. An MLST scheme for C. perfringens was developed by Xiao et al, comparing nucleic acid sequences of eight housekeeping genes of C. perfringens isolates and analyzing phylogenetic relationship between different species.¹² Two studies have shown that different toxin types of C. perfringens exhibited distinct genetic characteristics.^6,13

C. perfringens is well known in many countries as the causative agent of several forms of enteric disease;^4,14–16 however, the prevalence of enterotoxigenic C. perfringens isolates has not been systematically studied in hospitals in China. This study aimed to determine the prevalence, clinical characteristics, and molecular epidemiology of C. perfringens in hospitalized patients suffering from gastroenteritis, including foodborne GI diseases, AAD and SD, in the central region of China.

Materials and Methods

Sample and Clinical Data Collection

Fecal samples were collected from 1108 patients suffering from diarrhea and admitted to Henan Provincial People’s Hospital, Zhengzhou, China between Oct 2018 and Oct 2019. The samples were tested for C. difficile and C. perfringens. C. difficile toxin genes (tcdB, cdt, and tcdc deletion at nucleotide 117) from fecal samples were detected by GeneXpert C. difficile PCR assay (Cepheid Inc., USA). The available medical records of 112 patients with fecal samples positive for C. perfringens, were reviewed for age, gender, clinical profile, underlying disease, procedures performed and results of laboratory tests.

Culture and Identification of C. perfringens Isolates

The freshly collected stool specimens were delivered within 2 h of collection to a clinical microbiology laboratory to test for the presence of C. perfringens. After shock treatment with 95% alcohol, stool specimens were cultured anaerobically on 5% sheep blood agar plates at 37°C for 24 h. The colonies suspected to be belonging to C. perfringens were identified by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) (Bruker Daltonics GmabH, Billerica, MA, USA). Isolates were maintained in cooked-meat medium with glycerol (30%) at −80°C for further studies.

PCR Amplification of Toxin Genes

Bacterial genomic DNA was extracted from the colonies growing on 5% sheep agar plates using a Bacterial Genomic DNA Extraction Kit (Cat. No. 9763, Takara, Japan). Plasmid DNA was purified from the pure bacterial cultures using E.Z.N.A. Plasmid Mini Kit I (Cat. No. D6942, Omega, USA). The genomic and plasmid DNA were stored at −20°C till used for PCR experiments.

The toxin genes cpa, cpb, cpb2, etx, iap, cpe and netB were amplified by PCR using specific primer pairs^1,17 (Table 1). DNA amplification by PCR was carried out in a reaction volume of 25 μL with 3 μL of template DNA (200 ng/μL), 1.5 μL each of 10 pmol/μL forward and reverse primers, 6.5 μL of water, and 12.5 μL of 2× Multiplex PCR Mix. The PCR was performed using a Bio-Rad T100 system (USA) using the following program: Initial denaturation step of 5 min duration at 95°C, followed by 35 cycles of 30 s at 94°C, 1 min at 50 °C, and 1 min at 72°C. The final extension was performed for 10 min at 72°C. The PCR products (4 μL) were visualized by electrophoresis on a 2% agarose gel with a 2000-bp ladder (DL2000, Takara, Japan) as the molecular size marker. The type B isolate C. perfringens ATCC 3626 was used as a positive control for cpa, cpb and etx genes. Isolates SM101 and ATCC 13124 were used as positive and negative controls, respectively, for cpe gene.

Table 1 Primers Used for PCR Amplifications in This Study

MLST Analysis

MLST typing was performed as described previously.¹² In brief, eight loci (colA, groEL, sodA, plc, gyrB, sigK, pgk and nadA) were amplified by PCR and the PCR products were sequenced by using a 3730 XL DNA Analyzer (Applied Biosystems, USA). DNA sequences were submitted to a public C. perfringens MLST database (https://pubmlst.org/organisms/clostridium-perfringens) to obtain the sequences type (ST). New alleles and STs were deposited in the C. perfringens MLST database. The phylogenetic trees were constructed from the concatenated sequences by the maximum likelihood method, using a Tamura-Nei model in MEGA version 7.0 software.¹⁸

Statistical Analysis

Data were excluded from analyses if complete data sets were not obtained for a particular case. Statistical analyses were performed by IBM SPSS Software Version 20.0 (IBM, Armonk, NY). Clinical data were analyzed by the chi-square or Fisher’s exact test. A p-value of less than 0.05 was considered as statistically significant. Categorical variables were reported as frequencies and percentages, normally distributed continuous variables were reported as means and standard deviations (SDs) and non-normally distributed continuous variables as medians and interquartile ranges (IQRs).

Results

Prevalence of Toxin Genes and Toxin Types

A total of 153 C. perfringens isolates were collected from 1108 diarrheal stool samples. C. difficile pathogens were also detected in six specimens (3.9%) positive for C. perfringens. A total of 132 toxigenic C. difficile isolates were detected out of all stool samples by the GeneXpert C. difficile PCR assay. The cpa gene was detected in all isolates, cpe in 55.6% (85/153), cpb in 9.2% (14/153), and cpb2 in 3.3% (5/153) of the positive samples. The genes etx, iap, or netB were not detected in any of the isolates. C. perfringens type F (positive for cpa and cpe genes) was the main toxin type, accounting for 49.5% (75/153) of all isolates, followed by type A (cpa-positive), which was found in 38.6% (59/153) of the isolates. The type C isolates were grouped into two subtypes, type C-1 and type C-2, based on their toxin genes. Ten isolates carrying cpa, cpb and cpe genes were assigned to type C-1, and four isolates carrying cpa and cpb genes were assigned to type C-2. In addition, two isolates (positive for cpa and cpb2 genes) belonged to type H1, and three isolates (positive for cpa, cpe and cpb2 genes) were classified as type H2¹³ (Table 2).

Table 2 Prevalence of C. Perfringens Isolates and Toxin Gene Profiles

Association of Toxin Types of C. perfringens with GI Diseases

The 153 patients positive for C. perfringens infection suffered from three types of gastroenteritis (GI) diseases: foodborne GI diseases, AAD and SD. The food borne GI diseases were caused by type F (12/18), type C (3/18), type A (2/18) and type H2 (1/18) isolates. The 12 type F isolates could be further divided into five F^c (cpe gene on chromosome) and seven F^p (cpe gene on plasmid) types. AAD was mainly caused by type F^P isolates (30/48 cases), but a small number of patients were also found to be infected with type F^C (5/48), type C (5/48), type A (7/48) and type H2 (1/48). SD was caused mainly by types A (50/87), and F^p (23/87); however, a small number of patients were also found to be infected with type F^c (5/87), type C (6/87) and type H1 (3/87). Type F^P was found at higher frequency in AAD, compared to foodborne GI diseases and SD (62.5% versus 38.9% and 26.4%, p < 0.001), and type A was more prevalent in SD, compared to food borne GI diseases and AAD (57.5% versus 11.1% and 14.6%, p < 0.001). Type F^C had a higher rate in foodborne GI diseases, compared to AAD and SD (27.8% versus 10.4% and 5.7%, p < 0.001), There was no significant association between other types (type C, type H1 and type H2) and the three GI diseases (p > 0.05) (Table 3).

Table 3 Disease Caused by Different Toxin Types of C. Perfringens Isolates

Clinical Features of Toxin Types of Infected C. perfringens Patients

The data for age, gender, length of hospitalization, underlying diseases and results of laboratory investigations were obtained from the available medical records of 112 patients. C. perfringens infection was more common in the elderly patients and the patients older than 50 years were mainly infected with C. perfringens type F (39/60, 65.0%), type C (7/11, 63.6%) and type A (30/41, 73.2%). The age range of patients infected with type F (1–83 years) and type A (8–90 years) was broader as compared to that of type C (20–79 years). The ratio of male to female patients and the average period of hospitalization in type F, type C and type A infected patients were not different. Type F patients mainly suffered from malignancy (19/60, 31.7%), hepatobiliary disease (13/60, 21.7%) and ulcerative colitis (UC) (11/60, 18.3%). The main underlying diseases in type A patients were malignancy (16/41, 38.1%), hepatobiliary disease (10/41, 23.8%), cerebral infarction (7/41, 16.7%) and cardiovascular diseases (7/41, 16.7%). The numbers of type C patients were relatively small, the clinical course of type C patients was more severe. In type F and type C patients, C-reactive protein and Procalcitonin (ng/mL) were higher than type A patients (p < 0.05) (Table 4). Type C patients also had a higher incidence of fever (6/13, 54.5%), compared to type F (20/20, 30%) and type A 7/41, 16.6%) patients (p < 0.05). WBC count and fecal occult blood were similar in the three groups (Table 4). The mean length of hospitalization of patients suffering from foodborne GI disease was shorter (5.7± 2.3 days) than for those with AAD (34 ±15.6 days) or SD (11 ± 7.6 days) (p < 0.01). Similarly, patients suffering from foodborne GI disease and AAD had more serious C. perfringens infection and required a much longer duration of antibiotic treatment compared to those with SD (data not shown).

Table 4 Demographic and Clinical Characteristics in Different Toxin Types of C. Perfringens Groups

Molecular Subtyping and Phylogenetic Analysis

Genotyping analysis by MLST revealed that out of 153 C. perfringens isolates, 143 exhibited a wide variety of sequence types (STs), including many new ST types. Ten isolates could not be characterized due to double peaks in the sequences of PCR products which could not be resolved despite repeated experiments. The sequences of new ST types, corresponding to new isolates of C. perfringens had been submitted to C. perfringens MLST database (Table S1). A total of 83 isolates were chosen and analyzed genetically (Table 5), which included all type F and type C isolates, and 2 type A isolates (cpb2-positive isolates). The profiles of STs and toxin genes in the remaining type A isolates were shown in supplementary results (Table S2). MLST phylogenetic analysis indicated that type F^C isolates predominantly belonged to ST77 (6/15 isolates), which is more closely related to ST382, and the remaining 8 isolates grouped into a separate cluster. Six ST77 isolates were cultured from different wards and at different times, therefore ruling out the origination from a single source ward. In contrast, most of type F^P isolates exhibited considerable diversity, and only 5 out of 48 isolates belonged to ST41. Type F^P isolates were genetically divergent from type F^C isolates. Similarly, type C and type A carrying cpb2 gene isolates were assigned into different ST types respectively, and had high genetic diversity between ST types (Figure 1).

Table 5 Sequence Types (ST) and Toxin Gene Profiles of C. Perfringens Isolates

Figure 1 Phylogenetic analysis of several toxin types (Fc, Fp, C and A carrying cpb2 gene) of C. perfringens isolates created by the maximum likelihood method based on composite sequences of eight housekeeping gene fragments using MEGA 7.0 software.

Discussion

To our knowledge, this is the first in-depth investigation of fecal carriage of C. perfringens in clinical diarrheal patients in central China. We have reported the clinical features associated with infection by various subtypes of C. perfringens and molecular characterization and genetic diversity of various isolates. In this study, we found a high prevalence of C. perfringens type F^C infections in patients with foodborne GI diseases and C. perfringens type F^P were mostly associated with non-foodborne GI diseases. This observation is similar to recent reported studies.^4,19 Most C. perfringens type F isolates carrying cpe gene on the chromosome (F^C) caused food poisoning, some of them also caused AAD and SD. Likewise, most type F isolates carrying cpe gene on a plasmid (F^P) caused AAD, but a significant number of these isolates also caused food poisoning GI diseases and SD. Thus, C. perfringens type F isolates were a significant cause of AAD, food poisoning GI disease and SD diseases. Interestingly, we also observed that patients with AAD had longer hospital stays and duration of antibiotic treatments and more serious of C. perfringens infections. This finding supports a significantly pathogenic role of C. perfringens in AAD.

Beta toxin (encoded by cpb gene) is a pore-forming cytopathic toxin that can cause vascular necrosis, intestinal necrosis and systemic enterotoxaemia in humans and animals.^7,20 A recent case report described how C. perfringens type C triggered a life-threatening acute hemorrhagic necrotizing enteritis (AHNE) with high mortality.²¹ In the present study, we found that in spite of the small number of C. perfringens type C isolates cultured from these patients, infection with these isolates was characterized by vomiting, abdominal cramps, bloody diarrhea and even necrosis of intestinal mucosa. Some C. perfringens type C isolates (C-1) carried three toxin genes (cpa, cpb. and cpe). The increased virulence of type C isolates might be due to the synergistic effect of several toxins such as beta toxin with CPE. In this study, the clinical features and laboratory findings confirmed that type C isolates caused more serious clinical symptoms than type A isolates.

Beta2 toxin encoded by plasmid-borne cpb2 gene was associated with food poisoning and AAD. Five isolates carrying cpb2 gene were detected in diarrheal patients in this study. In a similar study performed in Japan, cpb2 gene prevalence was only present at a low rate.¹³ However, the C. perfringens subtypes containing cpb2 gene have been found to play an important role in pre-term necrotizing enterocolitis (NEC).^22,23 Additionally, cpb2 gene-positive C. perfringens isolates have also been found at higher rates in feces of autism spectrum disorder (ASD) children and infants aged <6 months.^24–26 Another study demonstrated that beta 2 toxin contributed to necrotizing soft tissue infections in hospitalized patients.²⁷ Further studies are required to investigate the pathogenic significance of beta 2 toxin produced by C. perfringens isolates in humans.

Patients older than 50 years-of-age were more likely to have C. perfringens infection. Similar results were reported in a study on C. perfringens infection in AAD patients from Iran.⁴ However, another study on C. perfringens infection in a community in England found an older average age for increased susceptibility.²⁸ The possible reasons that patients older than 50-years were prone to C. perfringens infection include three different aspects: (1) the majority of these patients have more comorbid diseases and weakened immune system function which cause them to be more susceptible to C. perfringens infection; (2) the diversity of intestinal microbiota in these patients decreases and loss of colonization resistance may result in an increased risk for developing C. perfringens infection after the use of antimicrobial therapy;²⁹ and (3) other therapeutic factors in these patients may impair colonization resistance, including surgery, cancer chemotherapy and invasive procedures that can lead to C. perfringens infection. The ratios of male to female and the length of hospital stay in the three groups of patients were similar to those reported in previous studies.^4,30 Typically, C. perfringens food poisoning is a self-limiting disease lasting 12–24 h; mortality is uncommon.³¹ However, we found that patients in our study positive for C. perfringens food poisoning experienced diarrhea lasting for more than 24 h, and had to be given antibiotic therapy during hospitalization. The most common underlying diseases in patients infected with both C. perfringens type F and type A, in decreasing order of prevalence, were cancers, especially GI tumors requiring chemotherapy and radiation therapy, hepatobiliary diseases, cerebral infarction and cardiovascular diseases. This finding is in agreement with the results of a previous study.⁴ Interestingly, we observed that patients with diseases of the hepatobiliary system such as decompensated liver cirrhosis and cholecystitis, were at greater risk of developing C. perfringens type F and type A infections. The main reasons are perhaps an imbalance in the enteric microbiota, alterations of the intestinal barrier, probably due to portal hypertension, and reticuloendothelial system dysfunction in these patients.^32,33 We found that a subset of patients infected with C. perfringens type F and A suffered from UC, confirmed by colonoscopy. These data are in disagreement with research by Aleksandra et al where the numbers of Crohn’s disease patients were higher than those of UC patients, but this may be attributed to the difference in the average age of the patients in these two studies.³⁴ The average age of patients in our study was much higher than the average age of 11.7 years in their study. It is important to emphasize the impact of C. perfringens type A isolates on patients with decompensated liver cirrhosis and UC patients with dysfunctional intestinal epithelial barriers. C. perfringens type A isolates can easily penetrate the destroyed intestinal epithelial barriers, enter the vasculature and cause bacteremia.

Based on the MLST scheme previously reported by Xiao et al,¹² we observed a considerable genetic diversity in the C. perfringens isolates. The phylogenetic analysis indicated that most C. perfringens type F^C isolates belonged to a single distinct cluster that evolved independently from C. perfringens type F^p isolates. Moreover, C. perfringens type F^p isolates do not have a common genetic background, which supports the idea of horizontal transfer of cpe gene among C. perfringens type F^P strains via conjugation (cpe-positive to cpe-negative transfer).³⁵ These results are similar to earlier studies that demonstrated a wide genetic diversity of C. perfringens type F^P isolates.^12,13 Our findings indicate that MLST assays can be used as tools to investigate the reservoirs and transmission among cpe-positive clinical isolates of C. perfringens.

The limitations of this study should be considered. Because there was not enough volume of stool specimens, we were only able to identify the toxigenic types of C. perfringens isolates using multiplex PCR, which is one of the diagnostic methods for ascertaining C. perfringens infection.^17,36 Indeed, several studies reported that C. perfringens infection was diagnosed by detection of CPE toxin using ELISA, reversed passive latex agglutination (RPLA) and enzyme immunoassay (EIA) tests.^28,37,38 Further experiments should be performed to detect the CPE toxin in stool specimens of patients.

Conclusions

We believe that testing for the cpe gene or CPE toxin should be included in the routine diagnosis of vulnerable patients over 50 years of age, or in patients with underlying diseases such as cancer, liver cirrhosis or ulcerative colitis. In addition, for patients younger than 50 years, if they are suspected of having food poisoning due to contaminated food, we also suggest screening them for cpe gene or CPE toxin. Lastly, our data add to the evidence base that despite its relatively low prevalence, C. perfringens type C can cause severe clinical symptoms in patients. Further studies are needed to carefully evaluate the potential role of C. perfringens type A in the etiology of diseases that involve violation of the intestinal barrier such as UC and spontaneous bacterial peritonitis in liver cirrhosis.