Infection through structured polymicrobial Gardnerella biofilms (StPM-GB)

Alexander Swidsinski, Vera Loening-Baucke1, Werner Mendling, Yvonne Dörffel, Johannes Schilling, Zaher Halwani, Xue-feng Jiang, Hans Verstraelen and Sonja Swidsinski

Charité Hospital, CCM, Laboratory for Molecular Genetics, Polymicrobial Infections and Bacterial Biofilms and Department of Medicine, Gastroenterology, Universitätsmedizin Berlin, Berlin, Germany, Vivantes Kliniken für Gynaekologie und Geburtshilfe, Am Urban and im Friedrichshain, Berlin, Germany , Outpatient Clinic, Charité Universitätsmedizin Berlin, Berlin, Germany, DRK Kliniken Berlin Westend, Klinik für Gynaekologie und Geburtshilfe, Berlin, Department of Obstetrics and Gynecology, the first affiliated hospital of Jinan University, China, Department of Obstetrics and Gynaecology, Ghent University Hospital, Ghent, Belgium and Labor Berlin, Department of Microbiology, Berlin, Germany


BACKGROUND: We analysed data on bacterial vaginosis (BV) contradicting the paradigm of mono-infection.

METHODOLOGY: Tissues and epithelial cells of vagina, uterus, fallopian tubes and perianal region were investigated using fluorescence in situ hybridization (FISH) in women with BV and controls.

RESULTS: Healthy vagina was free of biofilms. Prolific structured polymicrobial (StPM) Gardnerella-dominated biofilm characterised BV. The intact StPM-Gardnerella-biofilm enveloped desquamated vaginal/prepuce epithelial cells and was secreted with urine and sperma.

The disease involved both genders and occurred in pairs.

Children born to women with BV were negative.

Monotherapy with metronidazole, moxifloxacin or local antiseptics suppressed but often did not eradicate StPM-Gardnerella-biofilms. There was no BV without Gardnerella, but Gardnerella was not BV. Outside of StPM-biofilm, Gardnerella was also found in a subset of children and healthy adults, but was  ispersed, temporal and did not transform into StPM-Gardnerella-biofilm.

CONCLUSIONS: StPM-Gardnerella-biofilm is an infectious subject. The assembly of single players to StPM-Gardnerella-biofilm is a not trivial every day process, but probably an evolutionary event with a long history of growth, propagation and selection for viability and ability to reshape the environment. The evolutionary  memory is cemented in the structural differentiation of StPM-Gardnerella-biofilms and imparts them to resist previous and emerging challenges.


In 1683, Antoni van Leeuwenhoek sketched different morphotypes of “animalcules” in smears of dental film. Although he used other termini, rods, spirochetes, and cocci can be clearly recognized in his drawings. Leeuwenhoek’s discovery inspired research, which for the next 200 years is pointed to the appearance of microorganisms. Multiple “taxonomic” categories emerge, most of which were lost later and are unknown now. Ironically the name “bacterium” was introduced for non-relevant unclassifiable morphotypes. Despite a firework of microbial discoveries not a single pathogen was identified at that time. The idea of pathogens and likewise of antisepsis or asepsis simply did not fit into the existing logical framework. Microscopy in the first days disclosed complex microbial communities early everywhere. The abundance of bacteria, their ubiquity, polymorphism and readiness to emerge in organic substance made a practical link to involvement in disease difficult to imagine. The mainstream opinion was that bacteria can archigenetically uprise from nowhere (abiogenesis). Pasteur brought the  rchigenetic hypothesis down.

He demonstrated that all microorganisms are pre-existing, and need specific conditions for propagation, growth and transfection. The paradigm changed. Active isolation, transfection and culturing became standard procedures that identified the impact of microorganisms on disease, spoiling of food or  ermentation. The idea of intangible complexity was reduced to a handsome agglomerate of microorganisms, which can be approached one by one. The success  as overwhelming.

Due to isolation, a vast number of infectious diseases have been characterized and effectively controlled within the last 150 years. The emerging difficulties seemed to be temporary and solvable with development of new tools. The expectations remained unfulfilled. By the end of the twentieth century, it became increasingly apparent that the singularity is just one part of the story. As it is impossible to understand the human body by dissecting it into single cells, it is also impossible to deduce the functional potency of polymicrobial communities from the sum of involved bacteria. Bacteria exist mainly in the form of polymicrobial communities within the natural environment and thrive under conditions in which each of the community composing the microorganisms alone would perish. Microbiomes of the human body interfacing environment: skin, mouth, intestine, genitalia – are also polymicrobial. Infections such as  bacterial vaginosis, appendicitis, paradontosis, caries, tonsillitis, small bowel overgrowth, gallstones, inflammatory bowel disease, remain poorly understood,  espite 300 years of intensive research.

The vaginal microbiome is a typical example. Complexly composed, it shows remarkable constancy of its main players, which makes it unmistakably different from other microbial communities such as skin, mouth, and intestine. The sustainability of the vaginal microbiome is disturbed in conditions such as bacterial vaginosis, aerobic vaginitis, and fungal overgrowth. Bacterial vaginosis (BV) is the most common vaginal disorder. The disease was recognized in 1955 when Gardner and Dukes isolated Haemophilus vaginalis, now known as Gardnerella vaginalis from the vaginas of 92% of women with characteristic symptoms now known as bacterial vaginosis. In the same year, Gardner and Dukes were able to transmit BV using vaginal secretions of diseased women, but not when sing a pure culture of Gardnerella (Gardner and Dukes, 1955).

Criswell induced BV through inoculation with overnight cultures of Gardnerella (Criswell et al., 1969). Although these cultures could be impure, Koch’s criteria  seemed to be in general fulfilled: 1. isolation of pathogen from diseased but not from healthy persons, 2. cultivation of pathogen, 3. induction of infection by  ulture. The further progress came to a hold and even seemed to roll backwards. Investigators began to report up to 30-40% occurrence of G. vaginalis among subgroups of apparently healthy women, which made it difficult to accept G. vaginalis as a pathogen. At the same time, bacteria like Mobiluncus were isolated in BV, which were absent in healthy women. Unfortunately, the prevalence and persistence of these bacteria in BV was too low to explain the pathogenesis of the disease. The introduction of gene sequencing amplified the number of involved participants and further undermined understanding. More than 400 different acterial groups were allocated in vaginal smears of BV women. Since none of them fulfilled Koch’s criteria, bacterial vaginosis is traditionally explained by bacterial imbalance and overshadowed with lactobacilli by diverse gram labile microbiota. Unfortunately, the imbalance is a statement and not an explanation.

The widespread reasoning making a displacement of “good” lactobacilli responsible for increase of “bad” bacteria is even more ambiguous. Why do species,  which are ubiquitously present in the environment, suddenly vanish and do not recolonize the vaginal habitat? Why children, who have no stable indigenes, do not develop the disease at anytime, even when they are born to women with BV? Why is the incidence of BV not increased in the geriatric population or in newborns: groups that should be especially vulnerable to dysbiosis of any kind? Let us suppose the lack of lactobacilli is crucial. Could, for example, a lack of lactobacilli be restored (readmitted) by sex with a healthy partner? – Definitively no. Most studies show the sexual transmission of BV from diseased to healthy and not vice versa. Are changes of local pH, or substitution of lactobacilli curative? None of the previous studies could definitively demonstrate a sustained effect. Antibiotics, which in theory destabilize the vaginal microbiome, are the only substances with confirmed positive effects on BV. The history of bacterial vaginosis demonstrates perfectly the dead end of isolation based reductionism.

Vaginal bacteria are polymicrobial. To understand polymicrobials, we have to investigate them as a structure-functional unity. One helpful approach is the ribosomal-gene based fluorescence in situ hybridization or FISH. Each bacterium possesses 103-5 ribosomes.

Each ribosome includes a RNA copy. Some of the regions of the ribosomal RNA are strain-specific, others are universal for groups, domains, or even kingdoms.

Oligonucleotides, synthesized complimentary to sequences of interest and labelled with fluorescent dye, are called FISH probes. When added to samples containing bacteria, FISH probes hybridize with the RNA of the bacterial ribosomes. No additional enhancement of fluorescence is necessary and bacteria can be directly visualized because of the large number of ribosomes in each bacterium (Amann et al., 1995).

The following presentation summarizes our FISH studies of the vaginal microbiom. The names of the FISH probes are listed according to abbreviations of probeBase online resource for rRNA targeted oligonucleotid probes (Loy et al., 2007).

The results revealed by structure-functional investigations of vaginal microbiota strongly oppose accepted “facts”, often overturning traditional views. To demonstrate that the inconsistency stems from limitations of previously used methods, we introduce each result section with a short overview, explaining the history and origin of some (mis)interpretations.

In the last years a lot of literature has been dedicated to biofilms, each of the publications used often its own terminology, inconsistent with the terminology of other investigators. Before starting with the description we need to explain the terminology. Visible plaque, membranes and slimes coating the surface of human organs and prostheses are called films. In case of their colonization or contamination one supposes microbial biofilms. This assumption is often unjustified and leads to confusion in use of the term bacterial biofilm. For example “protective” vaginal lactobacilli biofilms, or “pathologic” Candida biofilms are often mentioned in the literature without any reference to their structure or to a single picture demonstrating their existence. Even we described lactobacilli in  time in our first publication mistakenly with the term “biofilm” (Swidsinski et al., 2005a). One should however distinguish between microorganisms suspended within secretions, which they do not build, and bacteria building their own matrix and growing in confluent layers. We will further use the word biofilm only in connection with self-maintaining biofilms creating their own environment.

Experimental procedures

The investigations were performed between 2004 up to the present. The materials investigated were: vaginal biopsies, desquamated vaginal epithelial cells from urine sediments, tissues from uterine curettage, surgically removed fallopian tubes, sections of nylon membrane strips placed into the vagina overnight, and sections of adhesive tape attached to the anal region for 60 seconds.

Fixation was performed with modified non-aqueous Carnoy solution (6/6/1 vol.  ethanol/glacial acetic acid/chloroform). The fixated materials could be stored in Carnoy at room temperature for up to 6 months. Usually, the time used was convenient for the investigated subject and laboratory staff.

Biopsies and tissue samples

Biopsies of about 3-5 mm diameter were taken from the middle sidewall of the vagina with biopsy forceps (Schubert, Aesculap, Tüttlingen, Germany, No. ER 058 R). They were immediately fixated in Carnoy solution for 2 hours. Surgically removed tissue samples were reduced to size of maximal 1×1 cm and fixated in 25 ml  of Carnoy solutions for at least 24 hours. Nylon membrane and adhesive tape strips were fixated overnight. Carnoy fixated material was processed and embedded into paraffin blocks using standard techniques. Four ?m sections were placed on SuperFrost slides (R. Langenbrinck, Emmendingen, Germany) or FISH studies (Swidsinski et al., 2005a).

Desquamated epithelial cells of the vagina or prepuce in urine sediments

Two ml of urine were mixed with 8 ml Carnoy fixative in 12 ml Falcon tubes. The collection and fixation was performed by the investigated subject. When possible, to increase the number of desquamated epithelial cells in urine sediments, women were asked to use the first portion of the morning urine and not to wash the region beginning with the evening before sample collection. Males were asked to pull the foreskin back over the glans penis before voiding.

An aliquot of 1.5 ml urine/Carnoy mix was centrifuged in a 1.5 ml Eppendorf tube for 6 minutes at 6,000 G. The sediment was decanted; the tube was filled with 1 ml of Carnoy solution and left at room temperature. After 1-5 minutes, the sediment was centrifuged once more (6 min/9,000 G), decanted, 50 ?l Carnoy solution was added, and then stored at 4°C. If more than 10 hybridizations were planned, larger urine aliquots were prepared in the same manner as described above.


Biopsies and tissue samples

Bacteria were assessed in a multi-colour analysis using a mix of specific and universal FISH probes stained with Cy3 (orange fluorescence), FITC (green fluorescence), Cy5 (dark red fluorescence) and DAPI counter stain (blue fluorescence) according to previously described protocols (Swidsinski et al.,  2005a; Swidsinski, 2006). GardV (Swidsinski et al., 2005a), Ato (Harmsen et al., 2000), Lab (Harmsen et al., 1999), Bac 303 (Manz et al., 1996), Ebac (Bohnert  t al., 2002) and Eub 338 (Amann et al., 1990) probes representing Gardnerella, Atopobium, Lactobacillus, Bacteroides/ Prevotella, and Enterobacteriaceae, Eubacteria cluster were applied for analysis of each sample. In special cases up to 140 further group-specific FISH probes were applied in different composed sets (Loy et al., 2007).

Nikon e600 fluorescence microscope, Nikon DXM1200F camera and accompanying software (Nikon, Tokyo, Japan) were used. The enumeration of bacteria was performed when hybridization signals were clear and morphologically distinguishable as bacterial cells by triple colour identification with universal and group-
specific FISH probes and DAPI stain, with absence of cross-hybridization with taxonomically unrelated probes (Swidsinski, 2006).

The conversion of the numbers within defined microscopic areas to concentrations of bacteria per ml was based on the calculation that a 10-?l sample with a cell concentration of 107 cells per ml has 40 cells per average microscopic field at a magnification of x1000, the details of conversion were previously described (Swidsinski et al., 2005a).

Urine sediments

A 5×5 mm quadrant area of hybridization was marked with a PAP pen on a SuperFrost plus glass slide. The Carnoy fixed urine sediment was vortexed; 5 ?l

aliquots were pipetted within the area of hybridization and dried for 30 minutes at 50°C just prior to the hybridization. Five microliters of the final aliquot were used for single hybridizations and represented 30 ?l of the initial urine volume.  Concentrations of epithelial cells within the 5×5 mm area of hybridization (30 ?l of sample volume) were calculated and converted to numbers of epithelial cells per ml of urine. The maximal and mean numbers of adherent bacteria per epithelial cell were determined.

The overall concentrations of adherent bacteria in the urine resulted from multiplication of mean number of bacteria per epithelial cell with the concentration of epithelial cells per ml of urine.


Our previous knowledge on bacteria attached to the vaginal epithelium is mainly based on indirect culture data, microscopy of vaginal secretions and swabs. Ridiculously, in the last 200 years no systematic study on bacterial colonization of human vaginal tissues was published and we do not know whether it was ever performed. We found no direct data in the literature except for singular microphotographs that illustrated viral, fungal or bacterial (toxic shock syndrome) infections and for electron microscopy of vaginal epithelium in animals. Contrary to the widespread belief of abundant bacteria covering the vaginal surface, the electron microscopy of healthy animals demonstrated, if at all, only attached single microorganisms. To clarify the real situation it was therefore  important for us to start specifically at this point.

Spatial organization of bacteria on the epithelial surface of the vagina

To investigate in situ the spatial organization of vaginal microbiota in different conditions 146 vaginal biopsies (Table 1) were obtained from women with BV, vaginal candidosis, aerobic vaginitis, and healthy pre- and postmenopausal adult subjects (cancer surveillance).

All women were gynaecologically examined and Gram stain of the vaginal smears was performed. Vaginal candidosis was diagnosed according to clinical symptoms and characteristic detections of yeast in the vaginal smear. Aerobic vaginitis was diagnosed in case of colpitis in absence of clue cells, lactobacilli and normal pH. The discrimination between BV and health was made according to the Amsel criteria after exclusion of vaginal candidosis.

Healthy women

No bacterial biofilms or any ordered bacterial adherence on the epithelial surface was seen in 54/56 healthy controls. Up to five singular lying bacteria of different bacterial groups could be seen irregularly scattered over the epithelial surface in some microscopic fields of 26/56 women, while most of their microscopic fields of view were completely free of bacteria (Fig. 1).

In 11/56 women islands of 5 to 20 clumped and irregular bacteria attached to the epithelial with a distance of one to two microscopic fields between each (300 ?m) were observed (Fig. 2).

Bacteria were in addition observed as clouds of diffusely scattered microorganisms in secretions covering the biopsy and not adherent to the epithelial surface in 17/56 women (Fig. 3, Table 1). The comparatively rare detection of bacteria in secretions covering the biopsy was in contradiction to the findings in the torn-out portion of all biopsies from healthy women, where bacteria were readily seen in considerable concentrations (10 to 50 per microscopic field) partially invading deeply the mechanically damaged tissues and suggesting that bacteria are present in the vaginal fluids and readily contaminate tissues as soon as their anatomic integrity is lost. However, the overall low occurrence of bacteria in secretions covering the vaginal epithelial surface may be due to difficulties to  reserve vaginal slime while processing biopsies in Carnoy fixative and in paraffin blocks, sections and FISH. Bacteria within secretions were more often seen in post- than in premenopausal women. Except for this finding, there was no noticeable difference in adherent microbiota between pre- and postmenopausal healthy women.

(…) the article as PDF file

Structured polymicrobial Gardnerella biofilms - as PDF file

Structured polymicrobial Gardnerella biofilms - as PDF file