Genotypic and phenotypic changes of Staphylococcus epidermidis during relapse episodes in prosthetic joint infections

Abstract

Staphylococcus epidermidis is a coagulase-negative bacterium capable of causing recurrent relapses in prosthetic joint infection (PJI). The aim of this study was to determine if Staphylococcus epidermidis isolates from patients with recurrent relapses of prosthetic joint infection (PJI) changed genotypically (pulsed-field gel electrophoresis (PFGE) pattern analysis and genes involved in biofilm formation) and phenotypically (antimicrobial resistance, biofilm formation) during the different episodes. Four patients with PJI recurrent relapses were evaluated clinically and microbiologically. Genotypic and phenotypic characteristics of 31 S. epidermidis isolates were determined. In all cases, PJI was treated with antimicrobial therapy and resection of the prosthesis without reimplantation. Months later, all patients had a relapse episode and treated with rifampin plus vancomycin and surgical debridement. Changes in the antibiotics resistance profile in isolates from patients 1 and 2 were observed in the two episodes. Patient 1 had four clones A, B, C, and D that were distributed differentially in the two episodes. Similarly, patients 2 and 3 had two clones and subclones (E-E1 and F-F1, respectively), and patient 4 had only the clone G in both episodes. The clone F formed small-colony variants (SCVs). High level of biofilm formation was found in all clones, except for clones D and G. Clones/subclones showed a genotypic variation in icaA, sdrF, bap, sesI, and embp genes. The principal coordinate analysis showed that all clones/subclones were different. These results showed that the initial infective clone of S. epidermidis from PJI, changed genotypically and phenotypically after a second relapse as a response to the treatment.

Introduction

Prosthetic joint surgery has a tremendous effect on the quality of life of the patients, and the demand for arthroplasty surgery continues to increase [1]. Nevertheless, not all prosthetic joint operations are successful because cases with post-surgery prosthetic joint infections (PJI) have been reported. The rate of PJI is low (1–2%), but when an infection occurs, it is often worrying for the patient and implies considerable costs for the healthcare system [1]. PJIs are classified according to the time of infection presentation; an early PJI is considered when it appears between 0 and 3 months after joint arthroplasty; a delayed-onset PJI occurs after 3 months but before 12 to 24 months, and a late-onset PJI occurrs from 12 to 24 months after surgery [2]. Early PJIs are caused mainly by high virulent microorganisms like Staphylococcus aureus and Gram-negative bacteria, while delayed-onset PJI and late-onset PJIs can be caused mainly by microorganisms with low virulence, e.g., Cutibacterium acnes and coagulase-negative staphylococcus (CoNS) [3, 4].

S. epidermidis is among the most important bacteria capable of causing late-onset PJIs, which is considered as a skin commensal bacterium, but currently, CoNS is the most prevalent in microbiological samples and the primary cause of CoNS-related infections, particularly in device-related infections such as prosthetic joints [5]. S. epidermidis can be pathogenic, especially when S. epidermidis isolates are biofilm producers and resistant to antibiotics [6, 7], which are the most common virulence factors in this bacterium.

Biofilm formation is the primary way by which S. epidermidis colonizes and infects prosthetic joints and medical devices. Biofilm is a structured community of bacteria attached to an inert surface or biological tissue, and bacteria are embedded in a mucous substance of self-produced extracellular polymeric matrix conferring protection against antimicrobial molecules (e.g., antibiotics, antiseptics) and immune response [8, 9]. The genes related to the biofilm formation are those that encode for cell surface proteins like SdrF, SdrG, Ebps, Bhp, and SesA-SesI [5, 10, 11], those proteins that participate in intercellular aggregation like Embp and Aap [5], or those proteins that participate in the production of the extracellular polymeric substances whose synthesis is mediated by the icaADBC operon [5]. Regarding the antimicrobial resistance, mecA gene is widely studied as a virulence factor because it participates in the inactivation of all β-lactam antibiotics [12].

Genotypic (pulsed-field gel electrophoresis (PFGE), MLST, genes related to biofilm formation) and phenotypic (biofilm formation and antimicrobial resistance) data show that the S. epidermidis isolates from PJI are different among them, suggesting that S. epidermidis PJI is diverse and have a high genetic diversity [13–15]. In the case of patients with late-onset PJIs and who had recurrent relapses, the bacterial clone that is infected for the first time is the same clone that causes recurrent relapses [16, 17]. There are currently few studies on the mechanisms of persistence and resistance to the antimicrobial treatments in patients with PJI who had recurrent relapses; we hypothesized that biofilm is an essential part of the mechanism of persistence during the transition from the first episode to the second PJI relapse. The present study was focused on determining if S. epidermidis isolates from a patient with recurrent relapses show genotypic change (PFGE pattern, presence of genes involved in biofilm formation) and phenotypic change (antimicrobial resistance, biofilm formation) during the different episodes.

Materials and methods

Study and patients

A cross-sectional study was carried out at the National Institute of Rehabilitation “Luis Guillermo Ibarra Ibarra (INR-LGII)” Mexico City, Mexico: a public teaching-hospital with 245 beds and orthopedic unit with 155 beds. The database of the infectology department from the INR-LGII was reviewed to identify patients that were diagnosed with PJI caused by S. epidermidis, between 2011 and 2017. S. epidermidis was found in 51 cases with PJIs; however, only 4 patients had late-onset and relapses caused by S. epidermidis. Medical records were reviewed and analyzed, and data are shown in Table 1.

Table 1.

Demographic information and clinical outcomes of the patients

Patient No.	Sex	Age (years)	Type of prosthesis	Prothesis age at time PJI (months)	Treatment	Clinical outcomes	Second episode of PJI (months)	Treatment	Clinical outcomes
1	M	54	Hip	30	(A) Antibiotics CC + RI (B) Debridament (C) Resection of prosthesis (D) Not reimplantation	Favorable	15	(A) Antibiotics RI + VA (B) Debridament	Favorable
2	M	62	Hip	139	(A) Antibiotics RI + VA (B) Debridament (C) Resection of prosthesis (D) Not reimplantation	Favorable	22	A) Antibiotics RI + VA (B) Debridament	Favorable
3	M	73	Knee	29	(A) Antibiotics RI + SXT (B) Debridament (C) Resection of prosthesis (D) Not reimplantation	Favorable	27	(A) Antibiotics RI + VA (B) Debridament	Favorable
4	M	41	Knee	33	(A) Antibiotics RI + LE (B) Debridament (C) Resection of prosthesis (D) Not reimplantation	Favorable	11	(A) Antibiotics RI + VA (B) Debridament	Favorable

CC+RIF clindamycin + rifampin, RI+VA rifampin + vancomycin, RI + SXT rifampin + trimethoprim-sulfamethoxazole, RI + LE rifampin + levofloxacin

This study was carried out following the recommendations of the ethics board committee of the “Escuela Nacional de Ciencias Biológicas-Instituto Politécnico Nacional (IPN).” The protocol SIP-IPN 20196630 was approved by the ethics board committee of the “Escuela Nacional de Ciencias Biológicas-IPN.” All subjects gave written informed consent under the Declaration of Helsinki.

Isolation and identification

The bacterial isolation was carried out by the hospital protocol for the diagnosis of PJI [18]. The procedure was as follows: the samples of periprosthetic joint tissues (soft tissues, bone, synovial fluid) or joint prosthesis collected during the surgery review were placed in Stuart Transport Medium (Becton and Dickinson, Franklin Lakes, NJ, USA) and taken to the laboratory. Soft tissues and bones were frozen in liquid nitrogen and then ground with mortar and pestle into a fine powder and later dissolved in 2 mL of sterile saline solution (0.85%). The joint prosthesis was placed in a sterile polypropylene container, and 400 mL of sterile saline solution (0.85%) was added. The solution was sonicated in an ultrasonic bath (Branson 3510, Mentor, OH, USA) at 40 kHz for 5 min. Aliquots of 0.1 mL of the homogenized and sonicated were spread onto 5% sheep blood agar, mannitol-salt agar, and phenyl-ethyl-alcohol agar. The Petri dishes were incubated at 37 °C for 48 h. Dishes with monoculture-like growth of S. epidermidis were selected for their bacterial identification. S. epidermidis identification was performed by the Vitek 2 computerized system (bioMérieux, Marcy L’Etoile, France). An infectious strain of S. epidermidis was considered when at least it grew in three samples of periprosthetic joint tissues (soft tissues, bone, synovial fluid) or the prosthesis [18].

Antimicrobial susceptibility

Antimicrobial susceptibility tests were carried out in the Vitek 2 computerized system using the sensitivity card AST-GP67. The antibiotics tested were as follows: oxacillin, gentamicin, ciprofloxacin, levofloxacin, erythromycin, clindamycin, linezolid, vancomycin, tetracycline, rifampin, and trimethoprim-sulfamethoxazole. All antimicrobial susceptibility results were analyzed and interpreted with the software Vitek 2 Systems (0.8.01; 2017) using the criteria of Clinical and Laboratory Standards Institute (CLSI 2017) [19]. Multi-resistant isolates were considered when they were resistant against at least three different families of antibiotics.

Determination of in vitro biofilm formation

Biofilm formation was performed in 96-well tissue culture plates (Nunc, Rochester, NY, USA) based on the method reported by Christensen et al. [20]. Bacteria were grown in individual wells of 96-well plates at 37 °C in tryptic soy broth (TSB; Becton Dickinson, NJ, USA) medium. After 24 h of growth, the plates were washed vigorously, dried for 30 min at 55 °C, and stained with 0.5% (w/v) crystal violet solution. The A_{492 nm} of the adhered, stained cells was measured in a Multiskan EX Microplate Photometer (Thermo Fisher Scientific, Lenexa, KS, USA). The criterion outlined by Christensen et al. [20] was used to determine whether isolates were non-adherent and biofilm-negative (A₄₉₂ < 0.12) or strongly biofilm-positive (A₄₉₂ > 0.12). Assays were repeated six times, and the mean biofilm absorbance values were used for the analysis. S. epidermidis RP62A was used as a biofilm producer and S. epidermidis ATCC12228 as a non-biofilm producer. The results were analyzed by one-way ANOVA with a Tukey’s test; the analysis was performed with the program GraphPad Prism version 5.0b

Genomic DNA extraction

Bacterial cells were grown overnight in TSB, harvested by centrifugation, and resuspended in 200 μL of lysis solution (20% sucrose, 10 mM Tris-HCl pH 8 and 10 μg/mL lysostaphin). Cells were incubated at 37 °C for 40 min, and 200 μL of Whinston solution (2% Triton X-100, 1% SDS, 10 nM NaCl, 10 mM Tris base pH 8.0 and 1 mM EDTA) was added. DNA was extracted with phenol/chloroform/isoamyl alcohol (25:24:1). DNA was subsequently precipitated with a volume of isopropanol and washed by the addition of two volumes of 70% ethanol. Finally, DNA was resuspended in sterile distilled water.

PCR amplification of genes related to the biofilm formation and mecA

Genomic DNA isolates were used as template for the amplification of icaA, ses, embp, and mecA genes, using primers previously described [21]. The PCR reactions were performed with 1 μL of DNA (100 ng), 1× buffer, 1 mM MgCl₂, 200 μM of each dNTPs, 1 U of Taq DNA polymerase (Invitrogen, Carlsbad, CA, USA), and 0.2 μM of each specific primer. The optimal PCR conditions were 30 cycles of 30 s at 92 °C, 40 s at 60 °C, and 30 s at 72 °C. The PCR products were analyzed on agarose gels. S. epidermidis IP34 was used as a positive control of amplification in all assays.

Pulsed-field gel electrophoresis

To evaluate the clonality, isolates from each patient were assayed. A PFGE protocol for S. aureus as described by The Centre for Disease Control and Prevention, Atlanta, USA, was established. The chromosomal DNA of each strain was extracted and digested with SmaI endonuclease (New England Biolabs, Beverly, MA). Restriction fragments were resolved in a CHEF-GenePath System (Bio-Rad, Hercules, CA) [22]. Classification of the clones was in accordance with criteria established by Tenover et al. [23].

Small-colony variants identification

Strains were grown on sheep blood agar for 24 h at 37 °C. Suspensions of each strain were adjusted to match the turbidity of a McFarland 0.5 standard, suspensions were diluted 1:1000, and 50 μL was spread plated aseptically on sheep blood agar for 24 h at 37 °C; three culture passes were realized to purify the small-colony variants (SCVs). The SCVs were defined in accordance with Bogut et al.; a SCV was considered as pinpoint colony, 1/10th or less than the size of normal staphylococcal colonies [24]. The SCV phenotype was confirmed by the auxotrophy test.

A test of auxotrophy for hemin, menadione, and thymidine was performed as reported by Bogut et al. with slight modifications [24]. Briefly, overnight cultures were diluted to match the turbidity of a McFarland 0.5 standard and inoculated on Mueller-Hinton agar (MHA; Becto Dikinson). Discs impregnated with 15 μL of menadione and thymidine at 100 μg/ml were placed on MHA. Hemin auxotrophy was investigated on MHA by using commercial hemin (X factor) discs (Becton Dikinson). Isolate was considered auxotrophic if it showed increased growth surrounding the impregnated disc.

Principal coordinate analysis

All genotypic (genes involved in the biofilm formation) and phenotypic (level of biofilm formation, antibiotic resistance profiles, and formation of SCVs) data were considered for the principal coordinate analysis. First, a presence or absence matrix of genotypic and phenotypic data from each isolate was constructed (9 clones/subclones). The matrix data were extrapolated to the program PAST 3.13 [25]. The proximity among isolates was determined by principal coordinate analysis by using the Jaccard similarity.

Results

Demography, clinical evolution, and treatment of patients with PJI

Four patients fulfill clinical and microbiological criteria for PJI. Patient demographic information, relapse dates, treatments, and clinical outcomes are shown in Table 1. All patients were men; the average age was 57 years; two had a hip prosthesis and two had a knee prosthesis. The average time between prosthesis implantation and first PJI was 58 months, which was considered as late-onset PJIs. The first PJI in all patients were identified by persisting pain and loss of the implant. In all cases, the first PJI was treated with antimicrobial therapy and resection of the prosthesis without reimplantation resulting a good response to the treatment (Table 1). However, all patients had relapses months later. In this case, the average time between the first PJI and the relapse was 19 months. In PJI relapse episodes, all patients were treated with rifampin plus vancomycin and they underwent to a surgical debridement.

Microbiological characteristic and clonality of the PJI isolates

A total of 31 S. epidermidis isolates of the two episodes from 4 patients were obtained from different body parts (Table 2). On average, three samples were cultured both first PJI and PJI relapse of each patient. The first PJI samples were bone, peri-prosthetic tissues, and soft tissue and were collected from peri-prosthetic tissues and soft tissue.

Table 2.

Antibiotic resistance and PFGE pattern in Staphylococcus epidermidis

Patient no.	Episodes	Date isolation	Source	OXA	GM	CI	LE	E	CC	LZD	VA	TE	RI	SXT	Clones/subclones
1	First PJI	14/10/13	Bone	R	R	S	S	R	S	S	S	S	S	R	A
		08/11/13	Soft tissue	R	R	S	S	R	S	S	S	S	S	R	A
		08/11/13	Synovial fluid	R	R	S	S	R	S	S	S	S	S	R	A
		08/11/13	Prosthesis	R	R	S	S	R	S	S	S	S	S	R	A
		08/11/13	Bone	R	R	S	S	R	S	S	S	S	S	R	A
		22/11/13	Soft tissue	R	R	S	S	R	S	S	S	S	S	R	A
	PJI relapse	25/11/15	Bone	R	R	S	S	R	S	S	S	S	S	R	A
		25/11/15	Soft tissue	R	R	S	S	S	S	S	S	S	S	R	B
		29/11/15	Soft tissue	R	R	S	S	S	S	S	S	S	S	R	C
		29/11/15	Soft tissue	R	R	S	S	S	S	S	S	S	S	R	D
2	First PJI	07/12/12	Bone	R	R	R	R	R	R	S	S	S	S	R	E
		07/12/12	Soft tissue	R	R	R	R	R	R	S	S	S	S	R	E1
		07/12/12	Prosthesis	R	R	R	R	R	R	S	S	S	S	R	E1
		17/01/13	Bone	R	R	R	R	R	R	S	S	S	S	R	E1
		17/01/13	Soft tissue	R	R	R	R	R	S	S	S	S	S	R	E
	PJI relapse	11/04/14	Bone	R	R	R	R	R	R	S	S	S	S	R	E1
		11/04/14	Bone	R	R	R	R	S	R	S	S	S	S	R	E
		11/04/14	Soft tissue	R	R	R	R	S	R	S	S	S	S	R	E1
		11/04/14	Soft tissue	R	R	R	R	S	R	S	S	S	S	R	E
3	First PJI	13/11/11	Bone	R	R	R	R	R	R	S	S	S	S	S	F
		13/11/11	Soft tissue	R	R	R	R	R	R	S	S	S	S	S	F1
		13/11/11	Prosthesis	R	R	R	R	R	R	S	S	S	S	S	F
	PJI relapse	16/05/12	Bone	R	R	R	R	R	R	S	S	S	S	S	F1
		16/05/12	Soft tissue	R	R	R	R	R	R	S	S	S	S	S	F1
		16/05/12	Soft tissue	R	R	R	R	R	R	S	S	S	S	S	F
4	First PJI	27/11/15	Soft tissue	R	R	S	S	R	R	S	S	R	S	S	G
		27/11/15	Prosthesis	R	R	S	S	R	R	S	S	R	S	S	G
		27/11/15	Bone	R	R	S	S	R	R	S	S	R	S	S	G
	PJI relapse	14/06/16	Bone	R	R	S	S	R	R	S	S	R	S	S	G
		14/06/16	Soft tissue	R	R	S	S	R	R	S	S	R	S	S	G
		14/06/16	Synovial fluid	R	R	S	S	R	R	S	S	R	S	S	G

A total of 31 S. epidermidis isolates from four patients were tested; they were distributed as follows: 10 from patient 1, 9 from patient 2, 6 from patient 3, and 6 from patient 4

OXA oxacillin, GM gentamicin, CI ciprofloxacin, LE levofloxacin, E erythromycin, CC clindamycin, LZD linezolid, VA vancomycin, TE tetracycline, RI rifampin, SXT trimethoprim-sulfamethoxazole

An antimicrobial resistance profile of 31 isolates was made. Patient 1 showed that 10 S. epidermidis isolates were resistant in both episodes to oxacillin, gentamicin, and trimethoprim-sulfamethoxazole. All 10 isolates had the same antimicrobial resistance profile, except for erythromycin where 6 isolates from the first PJI were resistant to erythromycin and 3 of 4 isolates from PJI relapse were sensitive to this antibiotic. The treatment for this patient included clindamycin and rifampin during the first PJI and vancomycin and rifampin during the PJI relapse, which was suitable for both episodes.

Patient 2 showed that the 9 isolates in both episodes were resistance against oxacillin, gentamicin, ciprofloxacin, levofloxacin, and trimethoprim-sulfamethoxazole. The isolates had changes in the antimicrobial resistance profile; 5 from the first PJI were resistant to erythromycin and 3 of 4 from PJI relapse were sensitive to this antibiotic; 1 of 4 from the first PJI was sensitive to clindamycin, but the rest were resistant in both episodes. The antimicrobial treatment was the same as in patient 1 giving similar satisfactory results.

Patients 3 and 4 did not show changes in the antimicrobial resistance profile in all isolates during both episodes. Six isolates of patient 3 were resistant to the following: oxacillin, gentamicin, ciprofloxacin, levofloxacin, erythromycin, and clindamycin. Six isolates of patient 4 were resistant to oxacillin, gentamicin, erythromycin, clindamycin, and tetracycline. The antimicrobial treatment was the same as in the previous ones.

Because the antimicrobial resistance profile of the isolates changed, it is possible that the isolates are different. In order to determine clonality, PFGE in all the isolates was performed; the PFGE patterns among patients were different; this suggests that the clones are different for each patient. In the case of patient 1, we found that 7 of the 10 isolates had the same PFGE pattern indicating it is the same clone which we call “clone A,” and 3 isolates had substantially different PFGE patterns indicating that they were new clones which we call B, C, and D. In patient 2, only 2 PFGE patterns were closely related in both episodes indicating that we have one more clone (clone E) as well as a subclone E1. This same situation occurred in patient 3, having the clone F and subclone F1. Finally, all 6 isolates from patient 4 had the same pattern indicating the clone G (Fig. 1).

Fig. 1.

PFGE patterns of clones and subclones of Staphylococcus epidermidis isolates. Samples: B bone, ST soft tissue, P prosthesis, SF synovial fluid

Considering the biological sample and episode and the clone/subclone distribution in patient 1 during the first PJI, all 6 isolates from samples of bone, soft tissue, synovial fluid, and prosthesis had the same clone A, suggesting a monoclonal infection; however, in PJI relapse, four isolates from bone and soft tissue had different clones (A, B, C, and D), suggesting a polyclonal infection and introducing new clones for this episode. The bone sample had the same clone A in both episodes indicating this clone persisted to the infection; however, soft tissue had clone A in the first PJI episode, while clones B, C, and D were found in PJI relapse episode, indicating a polyclonal change of infection for this sample.

Clone E and subclone E1 were found in the bone, prosthesis, and soft tissue from patient 2 during the first PJI; likewise, clone E and subclone E1 were found in the bone and soft tissue during the PJI relapse episode, indicating an infection by closely related clones and showing a persistence of these two clones in both episodes.

Clone F and subclone F1 were found in the bone, prosthesis, and soft tissue from patient 3 during the first PJI, but a change of clone F to subclone F1 was observed in the bone; similarly, a change of subclone F1 to clone F was observed in soft tissue during two episodes. This suggests a persistent infection by these types of clones/subclones in both episodes but there are changes of clones/subclones in the biological samples.

Finally, clone G was found in the bone, prosthesis, soft tissue, and synovial fluid from patient 4; similarly, this clone persisted in both episodes, suggesting a monoclonal infection (Table 2).

These results show that there are changes in the antimicrobial resistance and clones in both episodes of the same patient (mainly patient 1), demonstrating phenotypic and genotypic changes during the infection.

Biofilm formation and small-colony variants in isolates

Thirty-one isolates were analyzed for the presence of the genes involved in biofilm formation; all isolates from the same clone/subclone had the same genotype of genes involved in biofilm formation; for that reason, there is a correlation between clonality and genotype. Table 3 shows a list of genes involved in biofilm formation for each clone/subclone. Clones A, B, C, and D from patient 1 have different genotypes. Clone A has the gene sesI but not the gene sesA when compared with the other three clones; clone B differs from clone C by the presence of genes sesD (bap) and embp, and clone D differs from three clones by not having the genes icaA and sesD (bap).

Table 3.

Genes involved in the biofilm formation of Staphylococcus epidermidis isolates

Patient no.	mecA	IcaA	sdrF	sdrG	sesA	sesB	sesC	sesD (bap)	sesE	sesF (aap)	sesG	sesH	sesI	embp	Clones/subclones
1	+	+	−	+	-	+	+	+	+	+	−	+	+	+	A
	+	+	−	+	+	+	+	+	+	+	−	+	−	+	B
	+	+	−	+	+	+	+	−	+	+	−	+	−	−	C
	+	−	−	+	+	+	+	−	+	+	−	+	−	+	D
2	+	+	+	+	+	−	+	+	+	+	−	+	−	+	E
	+	+	-	+	−	+	+	−	+	+	−	+	−	+	E1
3	+	+	+	+	+	+	+	+	+	+	−	+	−	+	F
	+	+	−	+	+	+	+	+	+	+	−	+	−	+	F1
4	+	−	−	+	+	+	+	+	+	+	−	+	−	+	G

Clone E from patient 2 has sdrF, sesA, and sesD (bap) genes but not sesB gene when compared with the subclone E1. Clone F from patient 3 is different from subclone F1 by the presence of the sdrF gene. Clone G from patient 4 does not have icaA, sdrF, sesG, and sesI genes.

When determining the biofilm formation-level phenotype in the representative isolate of each clone/subclones, it can be observed that clones A, B, and C from patient 1 have high levels of biofilm formation and among them are similar (p > 0.05), but clone D has a significantly lower biofilm formation compared with the other three clones (p < 0.05). Patient 2 and 3 isolates had high biofilm formation levels and are similar between each clone of each patient (p > 0.05). However, clone G has a low level of biofilm formation similar to clone D from patient 1. The same biofilm formation level was observed in the same clone in different episodes (Fig. 2).

Fig. 2.

Levels of biofilm formation of Staphylococcus epidermidis clones. Biofilm formation was evaluated for each clone. 1 indicates first PJI, and 2 indicates PJI relapse episodies. (+) S. epidermidis RP62A was used as como control positive of biofilm formation and (−) S. epidermidis ATCC12228 as negative control of biofilm formation

In determining the formation of SCVs, clone F was the only one who formed SCVs, because it was stable after three culture passes (Fig. 3 (A–C)); it was an auxotroph for hemin, menadione, and thymidine and is slow-growing (Fig. 3(D)).

Fig. 3.

Small-colony variants (SCVs) of Staphylococcus epidermidis. A single colony of each clone was grown in sheep blood agar for 24 h at 37 °C (A). Later, a small-colony (B) and a large-colony (C) were grown in sheep blood agar after three culture passes. Growth curve of SCV and large-colony was realized (D)

These results show that clones/subclones are different in the presence or absence of the genes involved in biofilm formation, suggesting genotypic and phenotypic changes among them.

Principal coordinate analysis

By using the PFGE method, the presence of clones/subclones was determined in the patients. However, genotypic and phenotypic data indicate differences among them; therefore, to highlight the importance of genotypic and phenotypic data, a statistical analysis was performed. Principal coordinates analysis of clones and subclones shows that they are located at different coordinates, indicating that they do not share genotypic and phenotypic characteristics and are statistically different. In addition, this analysis suggests that the clones G and D are closely related to each other, followed by the clones A, B, and C and finally the clone E and subclone E1, similar to clone F and subclone F1 (Fig. 4). This result concludes statistically that the genotypic and phenotypic characteristics of clones/subclones are different, which makes it possible to discriminate among them.

Fig. 4.

Principal coordinate analysis (PCoA) of Staphylococcus epidermidis clones. All genotypic and phenotypic data were considered for the principal coordinate analysis and a data matrix of absence and presence was constructed. The analysis was performed with the program PAST 3.13

Discussion

The present study aimed to describe clinical and microbiological cases of late-onset PJIs relapsing caused by S. epidermidis and to determine genotypic and phenotypic changes in the clones during episodes of first PJI and PJI relapse. The retrospective analysis shows that the average time between the implantation surgery and the first PJI and the average time between the first PJI and the relapsed PJI correspond to late-onset PJI. To date, only two similar works to ours have been published; both reported similar times in the different episodes [16, 17]. Both works also mentioned that the clinical signs were as follows: implant loosening and persistent joint pain [16, 17]. In our study, all four patients had chronic joint pain and loosening of the implant. The patients were treated with antimicrobial therapy and resection of the prosthesis without reimplantation in first PJI, and PJI relapse consisted of surgical debridement and treatment with vancomycin and rifampin. This combination therapy is usually required for optimal management of PJI [26].

When comparing the antimicrobial susceptibility profile of the isolates against the two episodes, changes in antibiotic resistance was observed; this suggests phenotypic modifications of the isolates during the transition of two episodes; this is amazing because changes in antibiotics resistance during the episodes, especially erythromycin does not match to the antibiotics usually used in the treatment of patients. Nevertheless, this might suggest that modifications or transference of genes could occur in the clones as a response of survival to environmental conditions. These results were also reported by Henry et al. and Murillo et al. [16, 17]. On the other hand, isolates were considered as multi-drug-resistant and these isolates with this phenotype are considered hard to eradicate, as occurred with methicillin-resistant S. epidermidis in patients with PJI [27], indicating that multi-resistance is a virulence mechanism of this bacterium. Nevertheless, all bacterial isolates were sensitive to antibiotics used for the treatment in both episodes (first PJI and PJI relapse). Despite this, a second relapse occurred in four patients. This also has been reported in other types of infection; e.g., a patient with pacemaker-associated endocarditis caused by S. epidermidis developed break-through bacteremia despite taking antibiotics to which the S. epidermidis isolate is fully susceptible in vitro [28]. Considering all of the above, we can conclude that the mechanism of antibiotics resistance/multi-resistance is not a satisfactory explanation for persistent infection, suggesting that other mechanisms of protection in the isolates of S. epidermidis could also occur.

The PFGE patterns and principal coordinate analysis (PCoA) showed that each patient has their own clones/subclones, which are different and having a specificity of infection with genetic variation. Patient 1 was the only case that showed different clones in the two episodes, showing a monoclonal infection in the first PJI and polyclonal infection with new clones in PJI relapse episode. Henry et al. and Murillo et al. reported that isolates of each patient in recurrent infections were different strains; they defined this as “superinfection” [16, 17]. These results demonstrate that S. epidermidis can substantially change the structure of its genome to persist and cause relapses. In addition, it is demonstrated that the persistence of the infection can occur due to slight changes in the genome structure of S. epidermidis, since patients 2 and 3 had relapses and persistence of infection due to the presence of clones and subclones in the episodes. The presence of clones/subclones en PJI relapses has been not reported. Finally, we also found a persistent infection caused by a single clone, as in the case of patient 4. Similarly, Murillo et al. reported a persistent infection caused by the same clone in one of three cases [17], while Henry et al. reported the same in three of seven [16]. These results demonstrate a variation of events that may occur in PJI relapses due to persistent infection caused by S. epidermidis.

The persistence of foreign body–related infections have been associated in part to the pathogen’s ability to form adherence, biofilms on surfaces of implanted, or inserted medical devices [9, 29]. Biofilm formation was determined as a possible mechanism of persistent infection in all 4 patients. The different biofilm formation levels in the clones from patient 1 showed that S. epidermidis changed the biofilm formation level in the PJI relapse as a response to the infection process. This result is similar to that shown in the patient with pacemaker-associated endocarditis, where S. epidermidis increased the biofilm formation capacity and genetic diversification [28].

The differences in the biofilm formation levels are related to the presence of the genes involved in the biofilm formation, e.g., clones D and G which do not have the icaA gene. The icaADBC operon participates in the synthesis of the polysaccharide poly-N-acetyl glucosamine (PIA) [7], producing a biofilm more dense and robust that prevents the bacterial eradication with conventional antimicrobial therapy [9]. In this manner, all clones excepting clones D and G had an icaA-dependent biofilm as a mechanism of persistent infection. Clones D and G (without the icaA gene) could produce a biofilm less dense and robust, which is similar to Aap- and Embp-protein-dependent biofilm; this type of protein-dependent biofilm has been reported in clinical isolates de S. epidermidis [30–33]. Besides, we think that other different mechanisms to biofilm formation could be present in clones D and G, as well as in other clones; this mechanism might be the cell interiorization in tissues of patients to protect them from antibiotics. S. epidermidis causing of PJI treated with daptomycin and lysostaphin can interiorize inside the osteoblasts as a mechanism of persistent infection [34].

SCVs constitute a slow-growing subpopulation of bacteria with distinct phenotypical characteristics and pathogenic traits that facilitates persistent and recurrent infections [35]. SCVs of S. epidermidis have been identified as causative agent of persistent infections, because SCVs are recovered for several weeks or months from the implanted foreign bodies [36]; it is even assumed that both biofilm formation and SCV phenotype may contribute to the recurrence and persistence of PJI [37]. In our study, SCV was found in the clone F, suggesting that this clone could have another mechanism of persistence during the infection.

Another interesting result of our work is that the clones or subclones have different genotypes of the genes involved in the biofilm formation, demonstrating genetic changes of S. epidermidis during the infection. The different genotypes also suggest the mechanisms of biofilm formation in the isolates. The product of sdrG gene participates in the first step of biofilm formation (initial adhesion) [38]; all clones have the sdrG gene, which suggests that isolates start their biofilm by this mechanism. All clones have aap gene indicating this gene is important in its biofilm formation because it has been shown that the product of this gene participates in the second step of biofilm formation (the cellular aggregation) [39]. The product of the sesC gene is involved in the biofilm formation of S. epidermidis [40], the product of sesE gene suggests its participation in the cellular aggregation [41], and the product of sesH gene in S. epidermidis is unknown. sesA, sesB, sesD, and embp genes with high frequency in clones suggest participation in biofilm formation; however, the participation of sesA, sesB, and sesD genes in the S. epidermidis biofilm formation is unknown and the product of embp gene participates in the step of cellular aggregation [30].

We consider that the contribution of our work was to provide more information about the mechanism of persistent infection in the isolates of a patient with PJI relapse, focusing mainly on the changes of biofilm formation levels and the genes involved in its formation. This information is unknown in a PJI relapse, and the studies of Henry et al. and Murillo et al. lack this scope.

In conclusion, our results showed that isolates of S. epidermidis causing of PJI relapse produce genotypic and phenotypic changes in the first PJI and PJI relapse episodes as mechanisms of persistence and resistance towards the clinical treatment of patients.

Funding information

This work was supported by the “Consejo Nacional de Ciencia y Tecnología (CONACyT)”, Mexico, grant number 269242, and SIP-Instituto Politécnico Nacional Mexico 20181167 and 20196630. SOP received grant-aided support from CONACyT. SRM, MECD, and JCCD received support from the COFAA and EDI, Instituto Politécnico Nacional fellowships, and from Sistema Nacional de Investigadores, CONACyT.

Compliance with ethical standards

This study was carried out following the recommendations of the ethics board committee of the “Escuela Nacional de Ciencias Biológicas-Instituto Politécnico Nacional (IPN).” The protocol SIP-IPN 20196630 was approved by the ethics board committee of the “Escuela Nacional de Ciencias Biológicas-IPN.” All subjects gave written informed consent under the Declaration of Helsinki.

Conflict of interest

The authors declare that they have no conflict of interest.