Inflammatory processes caused by pathogens or by chemical or physical agents are important components in the pathogenesis of many human cancers (1). Increasing evidence suggests that also in the prostate gland inflammation may be implicated in oncogenesis. A recent study performed on 70,000 North American men showed that a clinical history of prostatitis significantly increases the relative risk for prostate cancer (PCa) (1.3; 95% CI: 1.1–1.5), and that a longer duration of the disease may further increase this risk (2). More indirect meta-analytic evidence showed that chronic consumption of aspirin or other NSAIDs can significantly decrease the odds of prostate cancer (PCa) (0.92; 95% CI: 0.86–0.97) (3).

In recent years, extensive research efforts have been devoted to investigate the link between inflammation and PCa and to find which inflammatory lesions of the glandular epithelium might act as early forerunners of neoplastic transformation. A new ‘injury-and-regeneration’ model, linking the effects of chronic inflammation to the molecular and cellular modifications underlying the pathogenesis of prostate cancer has been advanced by De Marzo and coworkers (4). According to this model, inflammatory cells infiltrating the prostate in response to an injury caused by infection or by the action of endogenous/exogenous irritants or cytotoxins may cause the initiation of neoplastic transformation by releasing genotoxic reactive oxygen species (ROS). Tissue injury and cell loss would at the same time trigger a tumor promotion step with proliferative regeneration of the damaged epithelium, leading to the appearance, mainly in the peripheral zone of the gland, of putative cancer ‘risk’ lesions of atrophic appearance, generally referred to as ‘Proliferative Inflammatory Atrophy’ (PIA). Proliferating atrophic cells, further exposed to ROS-induced oxidative DNA damage, might subsequently progress to the in situ cancer precursor prostatic intraepithelial neoplasia (PIN) and, ultimately, to frank adenocarcinoma (4).

The hypothesis pointing to PIA as a risk-lesion for prostate cancer has been investigated at the preclinical, clinical, cellular and molecular levels (reviewed in ref. 5). Very recently, the presence of atrophy has been linked to advanced prostate cancer (6,7). At the morphological level, it has been shown that high-grade (HG) PIN and PCa may merge with PIA lesions, thus suggesting the existence of a continuum between these entities (8,9). However, the issue of whether these histological findings are true signs of an undergoing PIA-cancer transition is still controversial (10–14), though some experts cautiously suggest that the great genetic instability of atrophic cells makes them more vulnerable to lesions possibly leading to neoplastic transformation (15,16).

At the genetic level, it was shown that a number of hallmarks of high-grade PIN and PCa are found in PIA cells (4,17). For example, chromosomal aberrations commonly found in PCa, like 8p22 loss, 8q24 gain, 8c gain and X gain, are also detected in PIA, albeit in the form of somatic aberrations (18–21). At the molecular level, a number of proto-oncogenes, tumor-suppressors and transducers of growth/survival signals such as NKX3.1, MSR1, Ki-67, Bcl-2, p16/CDKN2, p27, p53, GSTP1 and Cox-2, were found to be upregulated/disregulated in PIA, sometimes to an extent similar or identical to high-grade PIN or PCa (reviewed in refs. 4 and 5). The increasing evidence aimed at linking PIA, inflammation and PCa extends to findings involving the role of corpora amylacea (22), the expression of the prostate tumor overexpressed-1 (PTOV1) gene in atrophy (23), the involvement of immune regulatory cells (e.g., TH17 cells) and cytokines, and the influence of a number of PCa-associated polymorphisms in the cyclooxygenase-2 gene (reviewed in refs. 17 and 24).

Despite this mounting host of molecular evidence, very scant morphological/topographic data are available linking PIA or PIN to inflammatory findings or to clinical chronic prostatitis (CP).

To increase knowledge in this area, we investigated at the morphological level 1367 prostate biopsies from 98 patients affected by chronic prostatitis (CP) and 32 patients with a history of CP and a biopsy positive for carcinoma. The aim of this study was to investigate the topographic and quantitative relationship between inflammation, proliferative inflammatory atrophy and low- or high-grade proliferative intraepithelial neoplasia.

Equivalent proportions severe/moderate (50.28%) or mild/absent (49.72%) inflammatory infiltrates were detected in the biopsy cores analyzed in the present study. Similar numbers of atrophic lesions, dichotomized as PIA (n=840, 50.79%) or NON-PIA (n=814, 49.21%), were detected in prostate biopsies. Due to the equivalent proportions of PIA vs. NON-PIA findings, and of severe/moderate vs. absent/mild inflammation, normalization was not deemed as necessary. Low-grade and high-grade PIN were detected in total 307 and 6 biopsies, respectively.

Since its first definition, PIA has been considered as a lesion associated with, or caused by, inflammation. However, to the best of our knowledge, very few quantitative histological-clinical studies have been conducted to support this established empirical observation. The Billis and Magna and the Ruska et al studies are among the few published reports investigating the relationship between focal atrophy and prostate inflammation in autopsy or biopsy material (36,37). However, both studies focused exclusively on PIA and did not include NON-PIA lesions. Billis and Magna found that PIA was associated with inflammation in 66% of analyzed prostates, whereas in 22% of cases it was not (36). In our biopsy series simple atrophy, a PIA lesion, was associated with severe/moderate inflammation in 62.4% of biopsy cores (372/596 cores, Fig. 2), whereas in the remaining 37.6% of cases contiguity with inflammatory focuses was not evident. It must be stressed that we cannot exclude that in these latter cases an inflammatory focus might have been present in adjacent tissues, not included in the biopsy core. This is an important limitation of a biopsy study.

In contrast to the Billis and Magna study, in which PAH was exclusively associated with inflammation, we found that this atrophic lesion is more prevalent in biopsy cores with mild or no inflammation (63% of cases), and less frequently present in frank inflammatory specimens (37% of cores with severe/moderate inflammation) (Fig. 2). This latter percentage is strikingly similar to the fraction of PAH associated with moderate/severe chronic inflammation (34%) reported in the Ruska et al biopsy study (37). Thus, despite its inclusion into the ‘PIA’ category, PAH may appear to be in most cases a non-inflammatory lesion. Indeed, as stated by McNeal (38) and according to our experience, PAH may represent a ‘post-inflammatory’ atrophic lesion. This definition is supported by the morphological evidence that in PAH the glandular lumen often comprises cell-free areas [defined by Srigley as the ‘atrophic’ component of PAH, (39)], likely constituting the remnants of a previous inflammatory injury. Proliferative regeneration of secretory cells following disruption of the epithelium, may generate small, hyperplastic glands arranged in a lobular configuration, the typical feature of PAH. In our opinion, the fact that in over 60% of cases PAH was not associ ated with a moderate or severe inflammatory focus, is not sufficient to rule out the essential inflammatory nature of this lesion.

In the present study, partial atrophy, the most representative NON-PIA lesion, was found to be in a slight majority of cases associated with mild or no inflammation (56% of cores; Fig. 2). In our opinion, the fact that PA was found to be adjacent to moderate/severe inflammatory foci in 44% of analyzed cores should not be overlooked. Recently, Billis and coworkers have demonstrated that a mergence between PA and SA occurred in 27% of prostatic biopsies, with transitions often taking place within the same gland (40). The authors hypothesized that PA may be part of an evolving ‘continuum’ in focal prostatic atrophy, with SA arising from PA in this morphologic sequence. However, in no case did the authors find inflammatory cells in biopsy cores containing PA. Because of this finding, Billis and coworkers hypothesized that inflammation might represent a secondary phenomenon in complete focal atrophy. In contrast with these findings, in our study PA could be found adjacent to moderate or severe inflammation in more than 40% of cases, although in 217 cases out of 242 inflammation was classified as ‘moderate’ and in the remaining cases as ‘severe’. Thus, our data may in part support the hypothesis that PA could precede SA, as the result of a phenotypic sequential evolution, triggered by inflammation. This is supported by the fact that, like Billis and coworkers, we observed occasional PA-SA transitions within the same gland (data not shown). However, in contrast to Billis and coworkers, we maintain the view that inflammation is more likely to be a primum movens in the development of PIA, rather than a mere secondary phenomenon. A study is in progress to verify the ‘continuum’ hypothesis in patients subjected to multiple biopsies and/or radical prostatectomy.

An interesting finding in our study was the dramatic decline of the burden of low-grade PIN lesions, concomitant with the increase of the extent of PIA in the same gland. This inverse relation was best fitted by a power function decay curve (Fig. 6). This finding suggests that LG-PIN and PIA may be mutually exclusive lesions within the same tissue or tissue area, and it is supported by our finding that in about 95% of cases LG-PIN arises from a normal rather than from an atrophic gland (Fig. 5). In 2009, De Marzo and coworkers promoted their ‘injury-and-regeneration’ model for prostate carcinogenesis in the frame of a comprehensive review article (4). In that context, the authors included low-grade PIN as a possible intermediate lesion, developing from PIA, and representing a putative - although non-obligate - forerunner of high-grade PIN or cancer (Fig. 3 in ref. 4). The results of the present study seem to exclude this hypothesis, also given that low-grade PIN is more frequently found in non-inflammatory tissues (Fig. 5), which also seems incompatible with the higher prevalence of PIA adjacent to focuses of severe or moderate inflammation.

In the present study we failed to demonstrate a positive correlation between the extent of PIA lesions and the per patient burden of prostate cancer (Fig. 7). Intriguingly, prostate cancer and NON-PIA lesions tend to be mutually exclusive in the same tissue or tissue area (Fig. 7). This finding, to be confirmed in a larger sample of patients, weakens the hypothesis of a possible linkage between cancer and NON-PIA lesions. Notwithstanding this evidence, if a ‘continuum’ has to be hypothesized between partial atrophy, PIA and cancer, the former and the latter lesions are most likely to occur far apart in time.

Another interesting finding of the present study is the good correlation [Portney and Watkins criteria, (41)] between the total NIH-CPSI symptom score and the extent of both inflammation and PIA at diagnosis of chronic prostatitis (Fig. 8). This is to our knowledge the first study attempting to link the clinical symptoms of CP to the extent and severity of histological prostate inflammation, as well as to the extent of focal atrophy. The absence of a correlation between these entities at the end of therapy (few days before biopsy) is probably due to the attenuation of clinical symptoms of CP achieved by aggressive pharmacological therapy, and may indirectly support the link between inflammation and symptom severity during the active (inflammatory) phase of CP. Although the small sample size (n=34) does not allow to draw a conclusive answer, the correlation between NIH-CPSI scores and inflammation is indicative of a relationship between the clinical and histological aspects of chronic prostate inflammation. Whereas such a relationship can be expected in patients suffering from inflammatory subtypes of chronic prostatitis (class II CBP and class IIIa CP/CPPS), the correlation between symptom severity and the per patient extent of atrophy is more striking, and supports the hypothesis that an inflammatory injury caused by infection or other etiological determinants may be the original causative factor of the atrophic process (4).

In conclusion, the evidence emerging from the present study i) points to a positive association between tissue inflammation and PIA, ii) questions the presumed non-inflammatory nature of partial atrophy and iii) does not seem to support a model whereby low-grade PIN would arise from PIA lesions.