Introduction

In recent years, our understanding of the relationship between human-associated microbes and human health has undergone a profound change, with an increasing recognition that organisms not traditionally considered infectious in nature can contribute to disease processes. An organism’s pathogenic role can entail complex interactions with not only the host, but also the environment and other members of the human microbiome. This “pathobiome” model of disease challenges long-held notions of what defines a pathogen, obfuscates the classic metric of Koch’s postulates and raises entirely new questions about how microbe-associated conditions can be diagnosed and managed [1]. Falling into this exciting (but challenging) paradigm is the hypothesis that Propionibacterium acnes (P. acnes) can contribute to the etiology of various human disease states.

Propionibacterium acnes is a Gram-positive, aerotolerant anaerobe that forms part of the microbiota of the human skin [2, 3] as well as the oral cavity, large intestine and urinary tract [4]. It grows slowly under anaerobic conditions and has the capacity to produce biofilms in vitro and in vivo [4, 5]. P. acnes is routinely isolated from all patient specimen types submitted to the clinical laboratory except urine; its absence from urine cultures may reflect an inadequate incubation period [6]. Although P. acnes is a commensal organism important for skin health, it is also an opportunistic pathogen and the single dominant bacterium associated with the inflammatory skin condition acne after which it was named6. Its association with infections and clinical conditions beyond the skin has become increasingly recognized, particularly in relation to postoperative device-related infections [7], including those associated with prosthetic shoulder joints [8] and cerebrovascular [9], breast [10], spine [11] and cardiovascular device implants [12]. While a role for P. acnes biofilms in the pathogenesis of medical-device-related infections is now generally accepted [5, 13], there is also increasing evidence that the bacterium may act as an endogenous pathogen in a range of other human inflammatory conditions, particularly degenerative disc disease (DDD), raising the prospect of new diagnostic and therapeutic opportunities, including vaccination [14].

Recognition that P. acnes is associated with an increasing number of human conditions has always been tempered by the justified concern of contamination. Since P. acnes is part of the normal human microbiota, differentiation of true infection from contamination of biopsy samples can be a significant problem. As with other human commensals, Koch’s original postulates cannot be truly satisfied in relation to the establishment of a causal effect in disease. Therefore, other approaches must be pursued to establish a link or association. Conventional bacterial culture of potentially infected and negative control tissues is susceptible to false-positive results because it is sensitive to low-level bacterial contamination from the surgical environment, the clinical laboratory or the patient’s skin. It is also susceptible to false-negative results due to the presence of the biofilm matrix, which requires a rigorous disassembly step, such as tissue homogenization (when biofilm resides deep within tissue) [15] or sonication (when the biofilm resides on surface of the implant) [16] to enhance detection sensitivity, as well as its slow growth, which requires prolonged cultivation (for up to 14 days) [17]. Culture-independent molecular techniques for detecting P. acnes (e.g., direct tissue PCR) suffer from these same inherent limitations [15, 18].

As a result of these issues, conclusive demonstration of P. acnes infection or colonization requires a sensitive diagnostic technique, or series of techniques, that are conducive to the detection of true infection while eliminating the presence of contamination. Fortunately, imaging techniques such as fluorescence in situ hybridization coupled with confocal laser scanning microscopy (FISH-CLSM), immunofluorescence-CLSM (IF-CLSM) and immunohistochemistry (IHC) help fulfill these criteria by providing robust methods to visually identify biofilms that are insensitive to contaminating bacterial cells. The methods also provide definitive locations and morphologies of any P. acnes present. Bacteria with biofilm-like morphologies found deep within infected tissues cannot reasonably result from contamination since no time would be available for such matrix structures to form after surgical removal and immediate processing, or subsequent freezing pending analysis. Furthermore, and importantly, FISH-CLSM, IF-CLSM and IHC methods can also reduce or eliminate superficial contamination because sample preparation involves multiple harsh wash steps.

In preparation for this review, we surveyed the literature to identify human diseases associated with P. acnes infection, but for which definitive FISH-CLSM, IF-CLSM and IHC evidence of P. acnes biofilm within potentially infected human tissue from these conditions is available. In the following sections, we describe the current state-of-knowledge on the relationship between P. acnes and various diseases, including some not traditionally considered infectious in nature. These descriptions will include the underlying experimental methodologies used to generate the data. Although association does not indicate causation without further investigations, it does highlight how subtle host-microbe interactions with the pro-inflammatory P. acnes bacterium may be driving the pathogenesis of a number of diverse human diseases.

Study selection

Initial searches queried the PubMed database for original research articles published between September 17th, 1997 and August 30th, 2018. Articles published in any language were retrieved using the keyword Propionibacterium acnes AND/OR acne vulgaris, atherosclerosis, degenerative disc disease (DDD), prostate cancer, prostate hyperplasia, biofilm, infection, sarcoidosis and microscopy. A substantial proportion of the research articles initially retrieved using our search criteria were subsequently excluded (in vitro, cell lines, biology, genetics, genomics, animal models, treatment, reviews, commentaries, letters) because they did not describe studies of patients.

These initial searches yielded 2376 articles; 2033 of which were excluded using the search criteria described above. The remaining 343 original research articles were fully reviewed to identify the diagnostic methods utilized to detect the presence of P. acnes in human tissue. Of the 343 original research articles, 204 comprised retrospective or prospective studies and 139 were case studies (see Fig. 1).

Fig. 1
figure 1

Search criteria identified original research articles (September 17th, 1997 to August 30th, 2018) where tissue was assessed for the presence of P. acnes

Full size image
Fig. 2
figure 2

Search results identified original research articles (September 17th, 1997 to August 30th, 2018) where tissue was assessed for the presence of P. acnes

Full size image

The 204 remaining original articles where the tissue was assessed for the presence of P. acnes (case studies were excluded) were categorized into four groups: first-time original procedures utilizing tissue samples (n = 96), revision procedures with implant devices (n = 60), revision procedures involving tissues samples (n = 20) and samples from superficial swabs or surgical site infections (n = 28).

From the 204 original research articles and our exclusion and inclusion criteria, we identified 21 studies presenting microscopy-based evidence (e.g., FISH-CLSM, IF-CLSM and IHC with P. acnes-specific FISH probes or antibodies) for a clear role in human disease. These studies are presented in Table 1.

Table 1 Review of original research articles (9/17/97–8/30/18) where tissue was assessed utilizing FISH-CLSM, IF-CLSM and/or IHC
Full size table

Review of P. acnes biofilm infections in human disease

After carefully evaluating the studies presented in Table 1, and applying our strict inclusion and exclusion criteria, 10 studies remained that showed relevant microscopic findings (i.e., FISH-CLSM, IF-CLSM and IHC) regarding P. acnes infection in acne vulgaris, DDD, prostate disease and atherosclerosis.

Degenerative disc disease

Degenerative disc disease (DDD) is a multifactorial condition that can lead to pain and often will progress to degenerative osteochondrosis and other significant spinal pathology. Age and various environmental factors are known to increase risk, but no strong correlation has been shown [39]. A multitude of studies have revealed a significant heritability for disc disease, but the mode of inheritance is multifactorial and the high number of potentially causative genes limits their clinical applicability [40]. In addition to the mechanical and genetic factors that can lead to the pathogenesis of disc degradation, several studies have found evidence of P. acnes present in intervertebral disc tissue which may contribute to its deterioration [41, 42]. The majority of these studies, however, applied cultivation and/or PCR techniques, which are susceptible to the previously mentioned limitations of potential false-positive and false-negative results. A randomized controlled trial published in 2013 showed that a 90-day treatment with antibiotics significantly changed the clinical symptoms of DDD as well as the appearance of new Modic type-1 endplate changes when compared to placebo [43]. However, a fundamental weakness of this study is that the true presence of an infection was not investigated but implied based on the findings of another parallel study on a different group of discectomy patients. Furthermore, whether the observed difference between the antibiotic and the placebo groups could actually be attributed to the antimicrobial effects of treatment on a true disc infection or reflected a non-specific anti-inflammatory side-effect of the antibiotic employed is unclear. It would have been of great clinical value to verify whether certain cases of the degenerated discs were indeed infected.

Table 2 details all the studies related to DDD for which microscopic evidence is available and that meets our inclusion and exclusion criteria [i.e., (1) disc tissue that was formalin-fixed paraffin-embedded [FFPE] or cryopreserved; (2) a process that included harsh washing steps during sample preparation; (3) tissue sectioning; (4) use of P. acnes-specific FISH probes or antibodies].

Table 2 DDD original papers where tissue specimen was retrieved from first-time original surgeries/procedures and P. acnes biofilm found could not be explained by contamination
Full size table

In the first of two articles describing the visualization of P. acnes in DDD (see Figs. 3 and 4), the authors cultured bacteria from 162 (44%) of 368 disc specimens obtained during primary disc surgery for symptomatic disc herniations; of these, P. acnes was cultured from 119 (32%) [15]. Bacterial burden was assessed by determining the number of colony-forming units per gram of tissue. Isolates were subjected to rapid multiplex PCR phylotyping, and phylogroups IA1, IB, IC and II were observed. Eight culture-positive specimens were investigated using FISH, and P. acnes biofilms were observed in 7/8 specimens. Interestingly, infected patients were younger and more often male. This work led to the conclusion that a subset of herniated intervertebral discs are populated with P. acnes prior to any surgical procedure or intervention, and not, therefore, the result of contamination.

Fig. 3
figure 3

(Reprinted from Fig. 3, Capoor et al. [15])

FISH visualization of a P. acnes biofilm in disc tissue. a This color-combined image shows a “pocket” of green fluorescent P. acnes cells (biofilm) near the center right of the image. Red fluorescence from a general eubacterial probe (b) and green fluorescence from a P. acnes-specific probe (c). The b/c image is a zoom of a showing fluorescence from the red and green channels separately. Almost all of the cells in a are emitting both red and green fluorescence, indicating that they are P. acnes.

Full size image
Fig. 4
figure 4

(Reprinted from Fig. 2, Capoor et al. [15])

Visualization of bacterial biofilm in disc tissue. a Three-dimensional reconstructed confocal scanning laser microscopy image of P. acnes biofilm bacteria in a disc tissue sample (DNA stain SYTO9; green). b, c. The presence of P. acnes biofilms in this sample verified using fluorescence in situ hybridization (FISH). Epifluorescence micrographs of a biofilm cluster showing red fluorescence from the CY5-labeled EUB338 general eubacterial probe (b) and green fluorescence from a CY3-labled P. acnes-specific probe (c). Co-localization of the red and green fluorescence indicates that all of the bacteria in this biofilm were P. acnes.

Full size image

In the second of two articles describing the visualization of P. acnes in DDD (see Fig. 5), the authors sampled disc tissue from 51 patients surgically treated for a lumbar disc herniation, as well as 14 controls undergoing surgery for non-degenerative disease (most frequently fractures) [24]. As FISH-CLSM may, on rare occasions, produce artifacts that can resemble biofilms the authors imposed strict criteria for the classification of biofilm. These were (1) bacteria are organized in one or more aggregates, not as individual cells and should be embedded at least 2 μm within the tissue; (2) inflammatory cells are present in the tissue indicating a host immune response. Bacterial DNA was detected in 16/51 samples in the case group and 7/14 controls. FISH-CLSM demonstrated biofilm in 7/51 LDH patients and no controls. Interestingly, only one of these samples was PCR-positive for bacterial DNA. Furthermore, only 2/7 biofilm samples hybridized to the P. acnes probe. The strict biofilm-criteria used in the study are bound to produce false-negatives from the FISH-CLSM, as only a very small part of the disc is being examined. As such, the absolute numbers of positive samples are of little value. The main finding is that bacterial biofilms (including P. acnes) were clearly visualized, which supports the theory that a chronic bacterial infection may be closely associated with the development of some cases of DDD.

Fig. 5
figure 5

(Reprinted from Figs. 2, 3, 4 of Ohrt-Nissen et al. [24])

FISH of herniation samples, with bacterial aggregates hybridized to a universal bacterial probe (left) and a P. acnes probe (right, green color).

Full size image

Acne vulgaris

Acne vulgaris, known widely as acne, is a chronic inflammatory skin disease seen primarily on the face, neck, chest and back. While P. acnes is widely known for its association with this condition, its exact role in the pathophysiology of the disease is still unclear and controversial [44]. Currently, the standard view of acne is that androgen-induced hyperseborrhea and hyperkeratosis are thought to cause changes in the pilosebaceous unit that favor proliferation and dysbiosis of the P. acnes population [45, 46]. Androgen levels during puberty are increased which in turn activate sebaceous glands to produce more sebum which, along with hyperkeratinization, leads to clogging of the sebaceous duct and thus comedo formation. While the healthy sebaceous follicle appears to be colonized with a mixture of different P. acnes types, the microenvironment of the comedo, which is anaerobic with high lipid content, may possibly provide a growth advantage for “acneic strains from the type IA1 and IC phylogroups which subsequently flourish; strains from these phylogroups have been strongly associated with acne based on both metagenomics and culture-based studies [47].

Table 3 details the studies related to acne vulgaris for which microscopic evidence is available for a role in the condition, and that meets our inclusion and exclusion criteria described above (paragraph preceding Table 2).

Table 3 Acne vulgaris original papers where tissue specimens were retrieved from first-time original surgeries/procedures and P. acnes biofilm found could not be explained by contamination
Full size table

Jahns et al. [19] investigated the occurrence, localization and phylogroup of P. acnes in facial biopsies from 38 acne patients versus matched control biopsy samples obtained from patients undergoing benign nevus incision. To visualize P. acnes, they used an IFM protocol (with CLSM) with polyclonal and monoclonal (QUBPa3) antibodies targeting P. acnes, and monoclonal antibodies (QUBPa1; QUBPa2) that reacted with P. acnes type IA, type IC and type II. P. acnes was visualized in the sebaceous follicles of 48% patient’s samples versus 21% of the control samples, (p = 0.017). The P. acnes in the sebaceous units was found as either microcolonies of < 100 cells, some of which were attached to the hair shaft, or in extensive biofilms that penetrated deep into the sebaceous follicle. Biofilm was seen in 14 acne samples (37%) compared to five (13%) control samples.

Alexeyev et al. [20] and colleagues developed this previous investigation and determined the relative localization of P. acnes biofilm and CD3-positive inflammatory cells in the skin of 38 acne patients (median age 19 years) and 19 healthy subjects (median age 28 years). Within the patient group, eight comedones and 32 inflammatory lesions were analyzed. Facial punch biopsies were examined by IFM using the P. acnes monoclonal antibody QUBPa3, and via FISH using probes directed toward 23S rRNA and a putative lipase gene. IFM results demonstrated 18 patients (47%) positive for P. acnes compared to two controls (10%) (p = 0.0077). On the outer layer of the stratum corneum, P. acnes was found as isolated cells; however, within the stratum corneum P. acnes typically presented as microcolonies were sometimes matrix-encased. In the hair follicle, P. acnes were seen as microcolonies (> 50 bacteria) and macrocolonies (> 1000 bacteria). Finding large matrix-encased macrocolonies attached to the wall of the follicle was clear evidence of P. acnes colonization of the follicle. P. acnes macrocolonies/biofilm was found in 70% of colonized acne patients and 16% of comedones.

In Jahns et al. [21], a proof-of-concept study using a set of FISH and IF assays that allows simultaneous visualization of P. acnes and P. granulosum in skin biopsy samples was presented. The authors tested their IF assay on samples from six patients with acne vulgaris and two patients with benign nevus. Of the acne vulgaris patients, four patients tested positive for P. acnes, one of which also tested positive for P. granulosum. One of the samples revealed a combined biofilm of P. acnes and P. granulosum and the two organisms formed distinct populations. In a study from the same group, Jahns et al. [22] examined the distribution of P. acnes, P. granulosum and P. avidum in acneic and healthy skin by IFM and polyclonal antibodies targeting the cutaneous propionibacteria and found that P. acnes was the most common bacterial species in both control and acne samples, and about half of all colonized subjects showed evidence of biofilm in their hair follicles.

Atherosclerosis

Atherosclerosis is the major underlying cause of ischemia (restriction of the blood), leading to obstruction of peripheral arteries, congestive heart failure, heart attack and stroke. It is currently accepted that atherosclerosis is a gradual multifactorial process that occurs when low-density lipoprotein (LDL) and cholesterol from plasma collect beneath the endothelium (inner lining) of arterial walls following injury to the arterial lining. No consensus exists to indicate the prime cause of damage to the endothelial layer, but risk factors such as high blood pressure, diet, tobacco use, obesity and family history have been positively correlated with atherogenesis. The principal treatments for patients with atherosclerosis center around the reduction of LDL cholesterol in the blood, which is understood to be the principal determinant in the development of stenosis, or narrowing of the arteries, potentially leading to clot formation or rupture and blockage of blood flow to the heart or brain. Emerging research indicates that LDL cholesterol alone may not be responsible for atherosclerosis and, instead, there is strong evidence that other factors including inflammation and possible infection of the artery wall may play key roles in the onset and progression of arterial plaque deposition. Indeed, mounting evidence in the literature points to the regular involvement of bacteria within arterial plaque deposits including Chlamydia pneumonia and various Streptococcus and Staphylococcus species [48,49,50,51,52]. Furthermore, acute bacterial infections, especially of the respiratory system, have been demonstrated to often precede the onset of symptoms of stroke or myocardial infarction by several weeks. For instance, a strong, significant association between infections with a common pathogen of the upper respiratory tract, and cardiovascular disease has been shown [53].

Table 4 details the singular study related to atherosclerosis for which microscopic evidence is available and that meets our inclusion and exclusion criteria described above (paragraph preceding Table 2).

Table 4 Atherosclerosis paper where tissue specimen was retrieved from first-time original surgeries/procedures and P. acnes biofilm found could not be explained by contamination
Full size table

To date, we identified one study where P. acnes was found within the atherosclerotic arterial wall and could not be explained by contamination [23]. In this work, diseased human carotid arteries from 15 patients with advanced atherosclerosis were screened for the presence of Propionibacterium 16S rRNA, and an additional 10 arteries examined for P. acnes after culture in reinforced clostridial medium. DNA sequencing revealed 16s rDNA signatures of P. acnes in four samples, and nine samples were found to harbor a total of at least one or more species of cultivable bacteria, with 11 different species represented; in four of these cases, P. acnes was recovered. FISH was also performed on atherosclerotic carotid arteries from five additional patients to visualize the location of P. acnes within the plaque deposits, and to assess the relative contribution of P. acnes to the overall biofilm load within these infections. Detection of 16S rRNA targets was carried out using fluorescent-labeled peptide-nucleic acid probes (PNA-FISH) for total eubacteria and for P. acnes sequences in situ within 25 µm sections of carotid arterial plaque samples. Of the five arterial samples analyzed, all showed the presence of biofilm bacteria and four were positive for the presence of P. acnes. Bound probe was present throughout the 25 µm thickness of the tissue for all five samples analyzed and not solely on the surface, indicating the bacterial biofilms were attached and embedded throughout the tissue.

The four patient carotid artery samples that were positive for P. acnes demonstrated the location of eubacterial probe targets with green fluorescence and the location of P. acnes-specific 23S rRNA probe targets with red fluorescence. Each of the carotid arterial samples was found to be negative for the presence of bound nonsense-eubacterial 16S-rDNA probe. The four samples that were positive for P. acnes demonstrated that the probe locations aligned with other bacterial signatures in the sample, indicating a mixed biofilm. These findings showed that P. acnes was present and extensively involved within infecting biofilms in 4/5 atherosclerotic carotid arteries, that its involvement was typically associated with damaged tissue, and that it could be located throughout the media and into the adventitia. These observations also demonstrated that P. acnes was often associated with other unidentified eubacteria in polymicrobial biofilms.

Prostate cancer

An association between P. acnes colonization/infection of the prostate and malignancy was first proposed in 2005 based on microbiologic culture results from radical prostatectomy specimens of patients with prostate cancer [54]. Bacterial growth was noted in 19/34 specimens (56%), with P. acnes as the most frequently encountered species (observed from 15 patients). In subsequent years, other investigations have detected P. acnes, using a variety of experimental methods, from both prostatectomy specimens and biopsies of prostate tumors. These methods include: microbiologic culture [55, 56], P. acnes-targeted PCR [57, 58], pan-eubacterial 16s rDNA PCR with sequencing [57, 59] and non-targeted deep sequencing [60]. Across these studies, the detectable prevalence of P. acnes within tumor specimens has ranged significantly from ~ 23% to > 90% [55, 59], with other diverse bacterial species variably detected.

Perhaps the most convincing evidence for the presence of P. acnes within prostate tumors involves direct histologic observation. Visualization of the organisms in situ, with a defined topological/architectural relationship to the host tissue, rebuts the claim that P. acnes is an experimental contaminant. To date, several studies have utilized fluorescence microscopy in combination with organism-specific oligonucleotide probes or antibodies for the detection of P. acnes in prostate tissue.

Table 5 details the three studies related to prostate cancer for which microscopic evidence is available and that meets our inclusion and exclusion criteria described above (paragraph preceding Table 2).

Table 5 Prostate cancer original papers where tissue specimen was retrieved from first-time original surgeries/procedures and P. acnes biofilm found could not be explained by contamination
Full size table

Alexeyev et al. [27] described the visualization of P. acnes in diseased prostate tissue using a multicolor FISH assay that targeted P. acnes 23S rRNA. Using this assay, they demonstrated P. acnes infection of the prostate in 5/10 randomly selected prostate cancer patients, and sequential analysis suggested P. acnes infection can persist for up to six years. CSLM revealed intracellular localization and stromal biofilm-like aggregates as common forms of P. acnes infection in prostate tissues from both prostate cancer and BPH patients. Furthermore, in vitro infection studies with a prostate epithelial cell line revealed that type IB prostate isolate of P. acnes could form a biofilm-like conglomerate on the cells. The authors concluded that further study is needed to clarify the clear association between P. acnes, prostatic inflammation and the development of BPH and prostate cancer.

Studies by Fassi Fehri et al. [25] used in situ immunofluorescence and a polyclonal anti-P. acnes antibody to detect the bacterium in 58/71 cancerous prostate tissue specimens. The bacterium was not, however, present in 20 healthy prostates and 59 other cancerous tissue biopsies (mammary carcinoma). The bacterial load in each sample varied widely from a few cells to a large collection found almost exclusively in the secretory epithelium.

A more recent study by Kakegawa et al. [26] investigated the intracellular presence of P. acnes in longitudinal needle prostate biopsy specimens of patients with elevated prostate specific antigen (PSA) values. Detection was based on IHC and using a monoclonal antibody specific for P. acnes type I and II. The authors compared the presence of P. acnes in prostate tissue from 44 patients whose first biopsy specimens were cancer-negative, but ultimately progressed to cancer-positive, with 36 control patients whose biopsies remained cancer-negative throughout the course of the study. Although the number of P. acnes-positive macrophages did not differ between the groups, the percent gland-positivity was significantly higher in patients with cancer progression. Most notably, gland-positivity was higher both before the cancer diagnosis (i.e., at the initial negative biopsy 12.1% vs. 4.8%) and at the final biopsy (19.1% vs. 4.7%). These results suggest that the frequency of P. acnes-positive glands in the first negative prostate biopsy performed due to increased PSA titers can provide valuable clinical information for further biopsy or follow-up strategies.

Discussion

The studies presented in this review demonstrate the utility of FISH-CLSM, IF-CLSM and IHC in demonstrating the presence of P. acnes in diseased human tissue. Given the depth of the bacteria/biofilms within the tissues and the harsh washing steps inherent in these techniques, we are convinced that these findings represent true infection/colonization, and not instances of contamination. While the microscopy findings that implicate P. acnes in acne vulgaris may not be a surprise, similar findings in DDD, prostate disease and atherosclerosis support the emerging recognition of P. acnes as an opportunistic pathogen in other human diseases. Although association does not imply causation, it at least provides an initial platform for further evidence-based studies. Of particular relevance are the findings related to DDD since spine-related degeneration and its symptomatology are leading causes of disability world-wide; thus, treatments that prevent degeneration in even a portion of those affected would have a significant effect on the global population. Disc degeneration proceeds along a cascade that has been well defined using histopathological and radiological studies. The process of spine-related degeneration, which is thought to be multifactorial, has attracted a great deal of interest and research. While progress has been made, the search for all causative factors remains active. Well-established contributors to spine-related degeneration include biomechanical factors, changes in disc architecture with aging, traumatic factors and a genetic component; many of these are irreversible, making treatment algorithms for the prevention of disc degeneration based on such factors impractical. However, the identification of potentially reversible or treatable factors that affect the degenerative process represents tangible new directions in the search for expanded treatment options.

Propionibacterium acnes infection represents a new and compelling mechanism of disc degeneration that offers both a potential means of diagnosis and readily available treatment options. Early data support the presence of P. acnes within pathological disc material in patients with severe disc conditions requiring surgical treatment [24, 61]. It is important to define the role of P. acnes in the context of such discs, and to determine how the infection affects the degenerative process. Identification of P. acnes in degenerative discs could indicate that it promotes the degenerative process and, consequently, future treatment modalities for symptomatic disc degeneration could include treatment of the P. acnes infection. For example, P. acnes peptidoglycan has very pronounced stimulatory effects upon the immune system (e.g., Coparvax) and the presence of P. acnes as a biofilm within the discs and other tissues would act as a long-term local immune stimulant. A key element for the clinical application of antibacterial treatments is preoperative identification of potentially infected patients. Unfortunately, standard MRI has not been shown to correlate with intraoperative sample results. We encourage future studies to assess whether advanced imaging (e.g., PET-MRI) or biofilm biomarkers may play a role in this regard. Identifying a method for preoperative identification of infected patients will define the potential for antibiotic therapies for patients with symptoms of degenerative disc disease, or pathological intervertebral discs.

The role P. acnes appears to play in postoperative spinal infections is another important factor. The published literature indicates there may be two somewhat separate situations that we should consider in this context. First, there are low virulent and most frequently non-pyogenic implant infections whose most frequent common denominating symptom is local [62]. These low-grade infections tend to become clinically symptomatic years after the index surgery and might not be placed into context with a biofilm that originally was introduced during the index surgery, even when a culture positive for P. acnes is recognized as a real infection. This situation is similar to the one encountered with shoulder arthroplasty or ventriculoperitoneal shunts. Second, in spinal arthrodesis surgeries where the intervertebral disc is entered and the nucleus evacuated for an interbody arthrodesis procedure, resident P. acnes potentially contaminates the interbody implant and the fusion mass. This in turn might lead to so-called “failed fusions” and pseudarthrosis, as a recent publication on revision surgeries has suggested [63]. Spine surgeons may find these two quite realistic scenarios and their potential implications unsettling for several reasons. The thought that a hitherto unknown infectious variable with a potentially decisive influence on clinical outcomes might be at play in spinal surgery certainly is troubling. But the fact that we are not yet in a position to reliably test for the presence or absence of this variable makes such knowledge much more complicated. Should we routinely employ antiseptic rinses or topical antibiotics, which have their own risks and side effects, in certain types of spinal surgery? Which procedures should these be? What should be the consequence of a positive culture obtained during the index surgery, and should we routinely take samples for microbiology? Alternatively, do we need to perform a percutaneous disc biopsy prior to any interbody fusion for chronic low back pain in order to prove that P. acnes was preexistent in that particular disc and that the consequently higher risk for a suboptimal clinical result (a “failed fusion”), and need for closer postoperative follow-up and monitoring, was a preexistent risk and not imparted by the surgeon? The questions do not end here. The world-wide crisis of antibiotic-resistance prohibits us from prescribing antibiotics without a clear indication and without a precise rationale. Spinal surgery always has potential complications; postoperative spinal infection represents one of the most common and most-feared complication of spinal surgery. In particular, the frequency of postoperative P. acnes infections remains a significant problem. It is common to implant metal devices in the course of spinal surgery, and the presence of P. acnes in the discs presents a real risk of spinal implant colonization. The resulting postoperative infection can significantly decrease the success of the surgical procedure. The finding of P. acnes in disc material can account for the high number of postoperative infections with this organism. Preoperative treatment with specific antibiotics to eliminate this potential preoperative infection or colonization may be required to reduce the number of postoperative infections. Further study is needed to understand the role of P. acnes in the pathogenesis of disc degeneration, and to determine if preoperative protocols to address potential infection would reduce the occurrence of surgical complications.

Conclusion

Our strict study criteria of FISH-CLSM, IF-CLSM and IHC have provided solid evidence of P. acnes low-virulence (biofilm) infection in acne vulgaris, DDD, atherosclerosis and prostate cancer in first-time original surgeries. This, together with increasing evidence suggesting the presence of P. acnes as a pathogen in a wide range of human inflammatory conditions, suggests it is important for clinicians, surgeons and researchers to study diseases and infections with a suspected link to a P. acnes infection “under the microscope.”

Tables 6, 7, 8, 9, 10 and 11 classify 83 of the 96 first original cases (86%) from Table 2 according to 6 disease types [i.e., acne vulgaris (6), atherosclerosis (5), discectomy (26), prostate cancer (16), sarcoidosis (15), shoulder arthroscopy/plasty (15)]. These original cases highlight the possibility that colonization may occur through hematogenous spread following some mechanical damage, followed by local inflammation and healing.

Table 6 Review of first-time original procedure research (9/17/97–8/30/18) for Acne vulgaris
Full size table
Table 7 Review of first-time original procedure research (9/17/97–8/30/18) for atherosclerosis
Full size table
Table 8 Review of first-time original procedure research (9/17/97–8/30/18) for discectomy
Full size table
Table 9 Review of first-time original procedure research (9/17/97–8/30/18) for prostate cancer
Full size table
Table 10 Review of first-time original procedure research (9/17/97–8/30/18) for Sarcoidosis
Full size table
Table 11 Review of first-time original procedure research (9/17/97–8/30/18) for shoulder
Full size table

Against this context, it would be prudent to establish a biobank of surgical tissue and isolates taken from patients in whom P. acnes infection has been identified. Such biobanks would allow for future microscopy-based studies that provide additional confirmatory evidence of the extent of P. acnes infection and also facilitate in vitro and in vivo host-microbe investigations of DDD.

Looking forward, screening surgical tissue to determine the likelihood of an underlying P. acnes infection may provide valuable information for downstream patient treatment and management. Given its unusual anatomy, physiology and pathology, disc tissue is uniquely positioned to be at the forefront of such P. acnes low-virulence (biofilm) infection studies. The underlying studies highlight that P. acnes affects more than disc tissue and assists the spine surgeon in understanding intervertebral P. acnes infections are not an isolated phenomenon and underlines the importance of further research.