Identification of low oxygen-tolerating bacteria in prostate secretions of cancer patients and discussion of possible aetiological significance

The microaerophylic organism Propionibacterium acnes has shown consistent association with prostate cancer (PC). Studies linking circumcision with reduced PC further support anaerobes involvement as circumcision reduces anaerobe colonisation on the glans penis. A 1988 study linked anaerobes with PC but considered them as opportunists in necrotic tumour. A hypothesis that a “Helicobacter-like” process causes PC justified this pilot study. Active surveillance patients were enrolled. Post-prostate massage urine samples were screened using the Matrix-Assisted Laser Desorption/Ionization Time-of-Flight (MALDI-TOF) technique for bacterial identification after culture in anaerobic and aerobic conditions. 8 out of 18 patients (41%) had either obligate anaerobic (n = 5) or microaerophilic (n = 4, one of whom also had anaerobes) organisms identified. None of 10 control samples contained obligate anaerobes. Although mean PSA was 63% higher in those with low oxygen tolerating bacteria, two high outliers resulted in this difference being non-significant. Given the substantially higher proportion of PC patients with organisms growing in a low concentration of oxygen when combined with previous studies compared to controls, the degree of significance was as high as smoking 5–9 cigarettes a day and needs further investigation. Translational research in trials combining Vitamin D and aspirin have begun as part of such investigation.

There is considerable amount of evidence that chronic inflammation is a concomitant of cancer development in most sites, including prostate cancer 1 . It is unclear whether it is a precursor/promotor of cancer development 2,3 , just a manifestation of a response to the disorderly development that occurs as the cancer progresses 4 or a mixture of both as has been suggested by the hypothesis of cancer as a wound that does not heal 5,6 . Linkage between low oxygen-tolerating organisms and malignant transformation has been established with evidence for the association of the microaerophylic bacterium, Helicobacter.pylori, in stomach cancer 7,8 . More limited is evidence that Vitamin D deficiency is a promoting factor for this cancer type 9,10 . In prostate cancer there is no well accepted pathogen consistently associated with this malignancy nor the inflammation associated with it despite multiple studies, many of which have provided early suggestions but failed to be confirmed on subsequent studies 11 .
However, recent studies have suggested that infection with Propionibacterium.acnes (PA), a known sun-sensitive 12 microaerophylic bacterium associated with acne, has shown a more consistent association with prostate cancer [13][14][15][16][17][18] . Studies linking circumcision with reduced risk of prostate cancer [19][20][21] is further supportive evidence of the role of such organisms in prostate cancer aetiology as randomised trials of circumcision have demonstrated reduced anaerobe colonisation of the glans penis after circumcision 22 . The uncircumcised glans penis has direct access to the prostate through the urethra 23 which is relatively anoxic. Although there is one report from 1988 linking anaerobes with prostate cancer 24 and one study demonstrating that antibiotics can clear them from the prostate 25 , most workers have until recently considered them as opportunists colonising necrotic tumour tissue rather than being considered as actual causative factors. Due to a long-held view that a "Helicobacter-like" process could be involved in causation of prostate cancer [26][27][28][29] , this pilot study was undertaken to screen for aerobic and anaerobic bacteria by culturing small aliquots of urine samples obtained following a "mini-prostate massage" done for PCA3 testing. The use of a commercial MALDI-TOF (Matrix-Assisted Laser Desorption/Ionization Time-of-Flight) mass spectrometry device allowed for the rapid and accurate identification of bacteria.

Results
Clinical demographics of participants. Table 1 lists age, clinical and Gleason Grading pathology staging and PSA levels of 18 patients selected for study while undergoing a period of surveillance without having had any surgical, radiation or drug treatment. Table 2 lists the diagnosis in the group of non-prostate cancer controls undergoing microbiological urine testing for non-malignant causes who were selected for having adequate surplus urine for anaerobic culture.
PSA and type of bacterial isolate. Figure 1 shows results of dot plot distribution comparing the mean PSA of those urines containing obligate anaerobe or microaerophylic bacteria (mean PSA 11.7 range 4.5-26) and those with no such organism (mean PSA 7.2 range 1.29-12). Although the mean PSA level was 63% higher in the

Discussion
It has been known since the study of Cooper et al. 1988 24 that anaerobes are present more frequently in the tissue biopsies of malignant than benign prostates and that this frequency was higher than contemporary reports of urine cultures of normal subjects 30 . Recently there has been a cluster of reports showing a higher frequency of the microaerophylic organism P. acnes in malignant compared to benign prostates [13][14][15][16][17][18] and reports that circumcision both reduces anaerobe contamination of the glans 22 and reduces the risk of prostate cancer [19][20][21] . As a consequence, interest in both circumcision and these classes of bacteria has increased. This is particularly so since the cohort study of prostate cancer risk and circumcision by Spence et al. 20 . This study involved 1590 French Canadian residents and 1618 electoral registry controls. Although it failed to show any significant protection in the majority white population (OR 0.93 95% CI 0.80-1.12) except in men circumcised after the age 36 and non-significantly in the 26% of their men circumcised at birth (OR 0.86 95% CI 0.72-1.04), there was significant reduction (OR 0.44 95% CI 0.19-0.86) in the small subgroup of black men. More significant in respect of the results of our study was the fact that although there was no association in Spence et al's study with history of an STD, there was an association with a history of prostatitis. Perhaps the most important observation from this study was the observation that circumcision after the age of 36 gave protection. Clearly as an unpredicted observation it needs to be repeated but based on the observation contained in our report and the evidence on the role of phimosis in development of penile cancer 31 , the easiest way to recruit people to a study that clarified whether anaerobes do contribute to the raised PSA would be to target patients with raised PSA presenting with phimosis and measure PSA before and after surgery. Although our study is small and lacks adequate matched contemporaneous controls, these preliminary results certainly are different from those obtained from normal urine by others using earlier techniques 30 as well as our small series of "normal" controls, admittedly without prostate massage. However, pooling this limited literature data base of 27 prostate cancers and 551 "controls" (Table 4) produces a highly statistically significant difference (Pearson's χ² = 70.304 (df = 1, p < 2.2e-16) with Yates' continuity correction), providing justification for a prospective study including age-and sex-matched control samples from routine health care procedures such as cardiac stents and hip replacements and patients with benign prostatic hyperplasia (BPH).
In this small number of cases studied, no case was found with either P. acnes or H. pylori (the microaerophylic organism proven to have malignant association with stomach cancer 2 ) nor the facultative anaerobe, Streptoccocus gallolyticus, (associated with colon cancer 32,33 ) although there was one case (Tables 3, 4) associated with one of the Fusobacterium species, an anaerobe recently reported to be associated with pancreatic cancer 34 . Given the increasing numbers of cancers reported to have an association with anaerobes, they do give more support for the concept that, given the known hypoxic state of cancers, such organisms could indeed play a role either accelerating  Table 3. Oxygen dependence of organisms detected after prostate massage in study patients.
progression 35 or in addition playing a role in the induction as proven in stomach cancer. In this respect prostate cancer is possibly a particularly informative cancer to begin study of the mechanism involved given the data demonstrating reduced long-term sun exposure having a more clear-cut association (Table 5) with increased risk of cancer 28,36,37 , than spot Vitamin D levels 38 . As puberty is a time when males prove particularly sensitive to P. acnes and as with breast and cervix cancer there is some evidence linking early puberty with prostate cancer risk 27,39 , it could be that population-based studies of PSA levels in pubertal males 29 might give further evidence that could go some way to support the hypothetical interpretation of these data. This is that a 40-60 year life time of urban living with sub-clinical Vitamin D-induced, macrophage-mediated, non-antigen-specific immune deficiency 40 is as oncogenic in terms of the non-smoking related cancers as smoking between five to nine cigarettes a day over 40-60 years is a cause of lung cancer 41 (Table 5). An even more fruitful area for research is suggested by the observation of Head et al (Table 5) who demonstrated a 2.5 odds ratio for increased death from cancer at 10-15 years of follow up of civil servants taking sick leave for a psychiatric cause 42,43 . Many such people would be likely to rise with the sun, return home at sunset, take lunch at their desks and suffer from "non-seasonal-SAD" (seasonal affective disorder) because, whatever sun they were exposed to, did not follow the normal seasonal variation. Further population-based studies of "CBT-light" in such patients 44 could be very informative with serial monitoring of mood and anonymised screening of the bloods of recruited patients for PSA. Clearly such a study (limited to adults over the age of 40 given data that such individual with PSA greater than the median level of age corrected PSA had a 3.75 HR of developing prostate cancer 45 ) would only be justified if a non-surgical approach had been shown to normalise PSA long-term. Given laboratory studies in mice that demonstrated that this class of bacteria can be controlled therapeutically by treatment with Vitamin D 46 and the continuing amounts of data supporting the link between Vitamin D or sun exposure deficiency with increased risk of human cancer, trials of Vitamin D combined with anti-inflammatory agents such as aspirin aiming to reduce tumour progression (see http://www.isrctn.com/ISRCTN91422391 and https://www.provent.org.uk) could provide evidence to support the above interpretation. The increasing   recognition of the role of Vitamin D in enhancing non-antigen-specific macrophage anti-pathogen responses 40,47 provides an understandable mechanism. However, given the similarities in bacterial flora between what we have found and those in bacterial vaginosis, recently increasingly accepted as promoting HPV-induced cervix cancer (Table 5) 48,49 , it is likely for long-term control that treatment of both the patient and his partner may be necessary with selective use of appropriate anti-anaerobe antibiotic if response does not occur to Vitamin D alone. However, given the increasing problems with widespread antibiotic resistance and recent reports of successful use of E. coli vaccines in women with recurrent urinary tract infection 50 , it is likely that such a non-antibiotic approach may be preferable 51 .

No of cases studied Positive anaerobes
To conclude, MALDI-TOF enabled rapid and accurate identification of bacteria present in the post-prostate massage urine. A higher proportion of organisms growing in a low concentration of oxygen were found in prostate cancer samples than in previous normal studies and in the small number of "control" urines tested.
Given the association of cervical cancer with bacterial vaginosis and increased knowledge about how organisms such as Helicobacter pylori induce stomach cancer and association of Streptoccocus gallolyticus with colon cancer, larger scale studies in prostate cancer patients are required using techniques such as MALDI-TOF or Metagenomic sequencing 52 to rapidly identify bacteria which have hitherto been problematic to identify.
Given the increasing number of indications that the first event in the initiation of prostate cancer may be occurring in the early years of puberty, there is a need to find ways to better assess annual sun exposure to quantify the effect of the 30-40 years of chronic sun deficiency in modern urban environments. The students then undertook delivery to the laboratory and observed the screening procedure which was undertaken in microbiology in accordance with the approved protocol. In addition, a selection of surplus urines from routine urine requests from agematched men, were screened for obligate anaerobes only as well, although they had not had prostate massage.

Methods
Microbiology. 1.5 ml of urine was centrifuged at 16,000 g for 2 minutes. The supernatant was then discarded and 10 ul of the deposit plated on Columbia Horse Blood agar (Thermofisher, Basingstoke, UK) and Fastidious anaerobic agar (Thermofisher, Basingstoke, UK). Plates were incubated for 72h at 35˚C in air + 5% CO 2 or anaerobically in an atmosphere of 90% N2, 5% H2 and 5% CO2. Resultant colonies were identified by MALDI-TOF using a Bruker Maldi Biotyper mass spectrometer 53,54 . All data generated and analysed during this study are included in this published article.
Male death Lung cancer from smoking 5-9 cigarettes a day 59 1 11.1 (6.94-20.44) Table 5. Comparison of anaerobes and prostate/cervix ca risk and imprecise measures of sun exposure vs more precise lung ca risk of low dose tobacco.