Insulin-like growth factor 1 receptor affects the survival of primary prostate cancer patients depending on TMPRSS2-ERG status

Prostate cancer (PCa) is the most common cancer in men and the sixth cause of cancer death worldwide [1]. PCa is difficult to manage as it shows a spectrum of risk over time spanning from indolent tumors, which can be controlled with surgery or active surveillance, to tumors with aggressive and metastatic behavior that require more radical treatment strategies [2]. Consequently, there is an urgent clinical need for tools that can discriminate between the different conditions and stratify patients at diagnosis according to tumor progression risk. Established clinical and pathological prognostic factors, including serum PSA levels, Gleason score, lymph node involvement and the pathological stages of affected surgical margins, have proven useful but are insufficient for optimal risk stratification. From the genetic point of view, PCa can be considered a collection of cancers characterized by sets of molecular alterations that may underlie the clinically variable behavior of the disease and support the need to identify subgroups of patients with different prognoses [3]. Recently, the prognostic value of many molecular and genetic factors has been investigated, including the loss of PTEN or Akt mutations [4‐6]. The prognostic significance of the TMPRSS2-ERG (T2E) fusion gene, a specific chromosomal rearrangement found in 50–70% of PCa that involves the androgen-responsive promoter of TMPRSS2 and the ETS transcription factor family gene ERG, has been evaluated, but the results are not conclusive [7‐10]. The recent application of deep-sequencing techniques has led to a more comprehensive genomic portrait of localized and potentially curable PCa [11‐13], further pointing out the multifocal genetic nature of PCa and the presence of intra- and inter-tumor molecular heterogeneity that may affect tumor progression and response to therapy [14].

In the past years, several studies have recognized the prognostic role of some components of the insulin-like growth factor (IGF) system. The IGF system is composed of three receptors [insulin receptor (INSR), IGF-1 receptor (IGF-1R) and mannose 6-phosphate receptor (M6P/IGF-2R)], three ligands (insulin, IGF-1, IGF-2), and six known types of circulating IGF-binding proteins (IGFBP1–6) that modulate the bioavailability and bioactivity of the IGFs [15]. The IGF system has been reported to regulate normal and malignant growth, proliferation and differentiation, tissue homeostasis and cellular metabolism. The relevance of the IGF system and particularly IGF-1R in cancer has been widely documented [16]. The first evidence regarding the IGF system’s role in PCa came from epidemiological studies and showed that higher serum IGF-1 concentrations and decreased circulating IGFBP-3 are correlated with an increased risk of developing PCa [17]. In the prostate, IGF-1R plays a critical role in normal gland growth and development [18]. However, existing data regarding IGF system expression and its functional role in PCa are still controversial [19‐21]. Clinical studies evaluating the prognostic potential of IGF-1R are limited and report either positive or negative associations between IGF-1R expression levels and patient outcome [22, 23]. In this paper, we analyzed the expression of different components of the IGF system and their association with clinico-pathological parameters and the prognosis of biochemical progression-free survival (BPFS) and clinical progression-free survival (PFS) in a retrospective series of 270 patients with primary localized PCa treated with radical prostatectomy. In a previous study, we demonstrated that the IGF system is influenced by T2E as ERG directly binds the IGF-1R gene promoter, thus affecting its expression in PCa [24]. This paper shows for the first time that patients with PCa who do not harbor the T2E rearrangement and who express low levels of IGF-1R represent a subgroup of primary PCa tumors with poor outcome.

Although the relationship between the IGF axis and PCa risk and progression has been extensively studied, consensus is still needed. The discordance among studies is putatively due to different factors including i) composition of the analyzed series, ii) technical bias and iii) disregarded molecular mechanisms influencing IGF activity.

The findings reported in this study support a relationship between high IGF-1 and IGF-1R mRNA expression and favorable outcomes. Overall, the results are in contrast with the common view of IGF-1R as a marker of aggressiveness; however, previous studies of sarcomas and carcinomas reported similar results. In Ewing sarcoma, lower IGF-1 circulating levels were found in patients with metastatic disease [29], while in a cohort of 57 patients, a relationship between high IGF-1R and IGF-1 expression and favorable prognosis was found [30]. In breast cancer, lower expression of IGF-1R was found in tumor specimens than in matched control samples [31], and positive IGF-1R expression was associated with favorable prognosis [32]. Hence, our results are in line with evidence that the IGF-1R/IGF-1 axis is not an oncogenic driver in primary PCa. Plymate et al. demonstrated that restored expression of IGF-1R in malignant prostate cells slowed down growth both in vitro and in vivo [33], while an in vivo study by Sutherland et al. showed that conditional prostate-specific IGF-1R knockout caused cell proliferation, hyperplasia and the emergence of aggressive PCa when p53 activity was compromised [34]. Furthermore, Massoner et al. demonstrated that the IGF axis is up-regulated during normal epithelial differentiation in vitro [35]. In this study, we confirmed that up-regulation of IGF-1/IGF-1R signaling in local PCa is associated with a less aggressive phenotype.

Although recent advancements in next-generation sequencing technology have improved our understanding of the biology of prostate tumors [12], emphasizing the genetic basis of clinical variability of the disease, the impact of the molecular heterogeneity of PCa on the IGF axis has never been considered at clinical level. The genetic heterogeneity of PCa has become recently clear, and the molecular classification of PCa is helping to move towards a direct application of the personalized medicine concept. In this context, the presence of tumor-specific chromosomal translocations may have a crucial role. In 2005, Tomlins et al. first described the rearrangements of the ETS family of transcription factors (TMPRSS2-ERG) in approximately 50% of all PCa patients [36]. Since their discovery, these fusion genes have represented a powerful diagnostic biomarker. However, the prognostic significance of T2E is still controversial. Several authors have suggested an association between T2E and more aggressive tumor behavior and poor prognosis [37, 38]. In contrast, other studies have reported an association between T2E and favorable outcome, and still others did not find any association between T2E and patient survival [10, 39, 40]. In this study, we did not identify any prognostic relevance for the expression of T2E, a finding that is in line with a recent study that enrolled more than 1000 patients [41]. Nevertheless, when analyzing the expression of IGF system components according to the presence or absence of the T2E rearrangement, a difference in the value of IGF-1R expression as an indicator of disease progression was observed. Interestingly, these data were obtained not only dividing patients according to T2E status established using gold standard methods (PCR methods and/or FISH) but also as a result of ERG IHC evaluation. Accordingly to other studies [8, 42, 43], ERG IHC evaluation represents a reliable surrogate for T2E detection and is simpler and cheaper than molecular techniques. T2E was previously reported to influence the prognostic value of other genes. In this context, the prognostic value of SPOP was found to be statistically significant in the subgroup of patients not expressing the fusion gene [25]. In another study, high NBS1 gene expression was associated with BPFS in a subgroup of T2E-negative and PTEN non-deleted PCa patients [44]. The dependence of IGF-1R’s prognostic value on T2E status may partly explain the controversial evidence regarding the role of IGF-1R in PCa progression. The cellular genetic background may be relevant to modulate IGF-1R signaling and functions. In fact, IGF-1R functions may be affected by complex cross-talk with other signaling pathways and by direct interactions of IGF-1R with other cell surface receptors, such as the recently discovered connection with discoidin domain receptor 1 [45]. Several papers have demonstrated that aberrant expression of ERG alters the cellular transcriptional pattern, conferring a new phenotype characterized by an increased proliferation rate and/or invasiveness and decreased differentiation levels [46‐48]. The association between the IGF system and T2E rearrangement is not yet completely understood. The SWI/SNF chromatin remodeling complex, which high expression correlates with a prolonged disease-free survival in PCa patients, was demonstrated to both down-regulate transcription of TMPRSS2 and therefore the fusion gene and to sustain IGF-1 expression [49‐51]. Recent evidences demonstrated an interaction between the IGF system and T2E by identifying IGF-1R as a direct target of T2E [24, 52]. The reported clinical evidence indicates that T2E, with its wide spectrum of alterations, may counteract the putative beneficial effects conferred by IGF-1R expression, likely addressing cancer cells toward a less differentiated, more aggressive phenotype.

The results of this study provide new criteria for the classification of primary PCa patients based on contemporary assessment of T2E and quantification of IGF-1R expression. Particularly, the combination of an absence of T2E and low expression of IGF-1R identifies a group of patients with a poor prognosis who could benefit from a more severe treatment regimen. In addition, the data suggest an economic approach to patient stratification based on IHC ERG and IGF-1R evaluation. These results further support the importance of T2E for classifying the distinct biological entities associated with different risks of progression and prognosis. In conclusion, we believe these results provide a path toward more precisely establishing specific subtypes of PCa with distinct outcomes.