Introduction

Colorectal cancer (CRC) is one of the commonest cancers in Western society with a median 5-year survival of 10–15%, in patients with advanced disease [1]. Conventional chemotherapy and radiotherapy have improved survival in patients with advanced CRC. However, the risk of recurrence even after microscopically complete (R0) resection can still be as high as 50–75% in patients with advanced disease [2].

Molecular targeting strategies are of particular interest due to their potential to increase treatment specificity to the tumor.

Hormone therapy in particular, has revolutionized treatment of localized and metastatic breast, ovarian, endometrial and prostate cancer, which can be used in conjunction with other treatments or as a stand-alone therapy. The gastrointestinal tract (GIT) expresses more than 30 hormone genes, making it the largest endocrine organ in the body [3]. In fact, the first two hormones or ‘blood-borne chemical messengers’ were gastrin and secretin, initially described in the 1900s [3]. Despite this, there is currently no clinical use of hormones as treatment or as prognostic markers in CRC.

Our aim with this review is to summarise the literature on studies examining the potential of hormones as prognostic and/or therapeutic markers. We highlight the most promising ones in the hope that they can spur further research into their use.

Discussion

Adipose tissue

Obesity is considered a risk factor to cancer including CRC and hormones secreted from adipose tissue may exert a tumorigenic effect on the GIT [4]. There are two types of adipose tissue: the white adipose tissue (WAT) and the brown adipose tissue (BAT) [5]. The WAT is known to produce cytokines and hormones, whilst the BAT is highly metabolic and produces heat [5]. The adipokines, like adiponectin and lectin, are proteins with autocrine, paracrine and endocrine function and have been implicated in many diseases including cancer [5].

Adiponectin (APN)

Adiponectin is a 30 kDa complement C1q-related protein, which is mainly produced by the WAT [4, 6]. The APN gene is on chromosome 3q27 and can exist as low molecular weight (LMW) trimers, middle molecular weight (MMW) hexamers or high molecular weight (HMW) multimers before it is secreted [4]. There are three adiponectin receptors: AdipoR1, AdipoR2 and T-cadherin [4]. It is likely that the HMW isoform of APN is the most biologically active isoform. Both MMW and HMW tend to bind to the T-cadherin surface receptor which is linked to cell adhesion and communication via a calcium-dependent mechanism [4].

Low levels of circulating APN in the bloodstream is linked to obesity-related CRC [7]. APN is involved in many other signaling pathways e.g. mTOR, NK-kB, JNK and STAT3 [8] and it suppresses inflammation, insulin resistance, endothelial cell migration and adhesion [5].

The prognostic role of APN and its receptors has conflicting evidence in literature.

There seems to be a consensus that circulating adiponectin is negatively correlated to tumour stage [9, 10] and tumour grade [6]. A prospective study by Chong et al. [11] suggests that high circulating, pre-diagnosis APN carries an increased risk of CRC specific and overall mortality. This study was carried out on patients with CRC who provided blood samples prior to their diagnosis [11]. On the other hand, studies investigating tissue expression of APN and their receptors have not had consistent results. In a study of 58 patients with CRC and 30 controls, Ayyildiz et al. [12] found no influence of AdipoR1 and AdipoR2 expression on survival, although they did find a significantly reduced receptor expression in CRC patients. Sakellariou [13] and Gialamas [6] found that AdipoR1 is negatively associated with nodal stage and AdipoR2 is positively associated with TNM and shorter survival in lower stages of disease. In 2018, Chloe et al. showed that increased tissue AdipoR1 messenger ribonucleic acid (mRNA) is associated with increased mortality and shorter OS [14]. Conversely, Ayyildiz et al. [12] showed that tissue AdipoR1 and AdipoR2 have no influence on survival. However only 58 patients were recruited in this study and their body mass index (BMI) was not known [12]. Canhoroz [15] who measured tissue APN expression also showed that this is not associated with DFS. This study was also limited by the small number of patients recruited [15]. T-cadherin receptor which binds the biologically most active HMW isoform is less well studied but was investigated by Duan et al. [16]. He found that T-cadherin promoter methylation status was not associated with stage or lymph node status [16]. Further studies are therefore needed to validate this marker and its role in prognosis.

Adiponectin supplementation has shown promise in suppressing colorectal carcinogenesis, likely through the adiponectin stimulated protein kinase (AMPK) phosphorylation in both in vitro and animal studies [7, 17, 18]. AdipoRon, an APN-like synthetic molecule, which likely acts via the same pathway, has been shown to improve insulin resistance and glucose tolerance, but also has been shown to inhibit pancreatic cancer cell proliferation [19]. Its antiproliferative effects via the AdipoR1 and AdipoR2 receptors could be a potential chemoprevention therapy in CRC, although it still remains to be investigated [19].

Leptin

This is the most abundant adipokine and is involved in regulating hyperphagy, glucose homeostasis, growth, immune response and angiogenesis [20]. There are six isoforms of the leptin receptor and only two have been linked to intracellular signaling [20]. OB-Rb is the longest isoform of the leptin receptor and is normally produced in the hypothalamus where it controls food intake via appetite regulation [20].

Although leptin’s role in colonic tumorigenesis and proliferation has been well studied [21,22,23], only a few studies investigated its prognostic potential. Paik SS and colleagues showed that a low tissue leptin expression in CRC tissue samples is associated with more advanced disease whilst high tissue leptin expression is associated with a favorable OS and DFS [24]. Uddin et al. [25] showed that tissue Ob-R expression is associated with better OS. However not all studies agree. Wang et al. [26] found that leptin/Ob-R expression was associated with later-stage, easy metastasis and lower grade of disease. Jeong WK [27] also found that leptin expression was inversely associated with nodal stage but was not associated with differentiation, T stage and stage of disease. Guadagni et al. [28] looked at both serum leptin and adiponectin levels in patients with primary and metastatic CRC and compared it to controls. They showed that leptin levels were higher than in controls whilst serum ADN was lower than controls, concurring with previous studies [22, 29]. A low leptin/adiponectin ratio (L/A) was associated with an increased survival compared to patients with high L/A ratio [28]. In patients with metastatic CRC, overall survival was significantly greater than 90% in patients with a low L/A ratios compared to approximately 30% in patients with high L/A ratios [28].

Leptin inhibition resulted in a decrease in the growth of tumours in animal and human studies [30]. Although antibodies against leptin receptors have been developed, these have yet to be tested as a CRC treatment [31]. A leptin-derived peptide (L16) used to decorate the PEGylated liposomal doxorubicin (Doxil®) has increased Doxil’s therapeutic efficacy in a murine CRC model, thus showing promise as a future treatment to increase target efficacy to tumour cells [32].

Reproductive system

Oestrogen

Oestrogen’s role in CRC is unclear. It was first observed in 1969, when Fraumeni [33] noticed an increased risk of breast and colon cancer in a group of nuns who have a longer and uninterrupted exposure to oestrogen [33]. Conversely, premenopausal women and post-menopausal women on combined hormone replacement therapy, which increases the level of oestrogen, have a lower incidence of CRC compared to age-matched men, thus refuting the oestrogen link [34].

There are three forms of oestrogen receptors: ERα, ERβ and G-protein coupled oestrogen receptor (GPR30) [35]. The first two are nuclear hormone receptors [36]. Although both ERα and ERβ are expressed to various other tissues in varying degrees, the main ER subtype in normal colonic epithelium is ERβ [36]. This is progressively lost as malignancy progresses, whilst ERα is increasingly expressed with CRC development [37].

Some studies show that ERβ is inversely correlated to prognosis although not all agree with this finding [37, 38]. Liu et al. [39] found that higher GPR30 expression in CRC tissue is associated with a better survival compared to lower GPR30 expression. However, further study is required to identify whether GPR30 is upregulated or downregulated in different patient cohorts and whether this can be used as a prognostic or therapeutic marker.

Testosterone

Testosterone, like oestrogen, is a steroid hormone. It is responsible for the secondary male sex characteristics and spermatogenesis. Its role in colon carcinogenesis is unclear as some studies show that lack of testosterone has a protective effect [40], and others increase the CRC risk [41]. Normal colonic epithelial tissue expresses both androgen receptor A and B (AR-A and AR-B), whilst CRC only expresses AR-A and they bind dihydrotestosterone (DHT) [42]. A third type of membrane ARs has been described, which is a G-protein coupled receptor and only found in colonic cancer cells (CaC02 and HCT116 cell lines) but not on normal colon cells [43]. Colon cancer tissue from xenografted tumours in mice, as well as the above cancer cells were used to show mAR-mediated tumour regression, thus showing promise as a novel therapeutic target [43].

Androgen’s capacity as a prognostic factor in CRC has not been explored in literature as far as these authors have found.

Pancreas and liver

Insulin and Insulin-like growth factor (IGF)

Epidemiological studies have linked obesity with an increased risk of all cancers including CRC. Insulin resistance seems to play an important role and chronic hyperinsulinaemia, which is associated with type 2 diabetes mellitus (T2DM), is linked to increased levels of insulin-like growth factor binding protein 1 (IGFBP-1) and IGFBP-2. This in turn leads to increased IGF-1 which promotes cell proliferation [44]. There are six high-affinity-binding proteins (IGFBP-1 to IGFBP-6) in the circulation, which carry the IGF-1 and -2, and are readily measured in the circulation [44].

Higher IGFBP-2 or low IGFBP-3 plasma levels were associated with worse OS in CRC patients [45]. High levels of circulating IGF-1 or IGF-2 as well as high IGFR-1 activity have been linked to early relapse, independent of KRAS status [46].

Drugs targeting the insulin-IGF pathway have been under investigation: dalotuzumab, a humanized IgG1 anti-IGR1 antibody by Merck, has been tested as adjunct therapy to cetuximab-based therapy for chemorefractory, KRAS exon 2 mutant metastatic CRC case with no improvement in outcome [47]. Use of NT157, an inhibitor of IGF-1 receptor and STAT3, in APC mice, has been shown to reduce tumour burden and metastatic potential [48]. Finally IGF-1 receptor tyrosine kinase inhibitors (e.g. AXL, NVP-AEW541) have been investigated in animal and in in-vitro studies and have shown antineoplastic effects, especially when used with 5-fluorouracil or cetuximab or fluvastatin [49, 50]. These need further study to assess their clinical applicability to humans.

Gastrointestinal tract

Pro-gastrin (PG)/gastrin (G)/gastrin-releasing peptide (GRP)

Gastrin was initially discovered in 1905 as a major stimulant of acid secretion [51]. Gastrin releasing peptide (GRP) which is released from the vagus nerve, stimulates the release of gastrin from the G cells of the stomach. Gastric acid is secreted when gastrin binds to CCKBR receptors on enterochromaffin cells (ECL) and parietal cells. When the stomach pH decreases, the antral D cells release somatostatin, which in turn inhibits gastrin release from G cells [51]. The main products of the gastrin gene are the amidated forms of gastrin 17 and gastrin 34 (G17-NH2 and G34-NH2). In CRC, the non-amidated forms of gastrin are predominantly expressed. However, many studies investigating the link between gastrin and CRC, seem to measure only the amidated forms, which may be the reason for the conflicting evidence on the role of gastrin in CRC carcinogenesis [51]. It is also postulated that cancer cells lack the enzyme to convert PG to amidated forms of gastrin and as a result, cancer cells mainly produce PG [52].

PG which binds to annexin II (ANXII) receptor which is overexpressed in human CRC, can promote mitogenesis through various signaling pathways (nuclear factor-kB, Janus-activated kinase 2, signal transducer and activator of transcription 3, extracellular signal-regulated kinase, mitogen-activated protein kinase and phosphatidylinositol 3-kinase/Akt kinase) which can enhance tumour growth [53]. PG overexpression in patients with hyperplastic polyps showed a 5-year neoplasm-free survival rate of 38% compared to 100% patients who had none or low progastrin expression [54].

Najib et al. [52] have recently shown a role of PG in promoting angiogenesis in in vitro and in vivo conditions. Knocking down of PG expression in the tumour cells by shRNA has shown a significant decrease in vessel density, making PG an attractive therapeutic target in CRC [52].

Gastrin-releasing peptide (GRP) is a neuropeptide that, as the name suggests, stimulates gastrin release. GRP and its receptor (GRPR) are not normally expressed in colonic epithelial cells but seem to be overexpressed in colon, breast, pancreas and lung cancers [55]. Rivera et al. [55] showed that high expression of GRPR alone or both GRP and GRPR, are associated with delayed recurrence (14.1–17 months respectively, p = 0.005) and increased survival (10.1–13.1 months respectively, p = 0.0124), as well as fewer lymph node metastases, making it an attractive prognostic marker [55]. Bombesin analagues (which bind to GRPR) have been used in radiolabelled form for imaging diagnosis, staging, recurrence detection and assessment of mestastasis in prostate cancer patients but have yet to be explored in CRC [56].

Cholecystokinin (CCK)

Cholecystokinin is produced by I cells of the proximal small bowel and has a role in digestion, appetite control and body weight regulation [57]. CCK binds to both of its G protein-coupled receptors (CCK1R and CCK2R) with high affinity. CCKR activation is implicated in proliferation, migration, differentiation, anti-apoptosis and inflammation and a single-nucleotide polymorphism in the CCKR gene has a prognostic potential in pancreatic cancer but not in CRC [58].

A recombinant immunotoxin (rCCK8PE38), which targets CCK2R, has been tested in 2 colon cancer cell lines (HCT-8, SW116) and one gastric cell line (MKN45) [58]. When nude mice were inoculated with HCT-8 and then treated with this immunotoxin, tumour regression occurred, unlike the negative control group [58]. The exact mechanism of action CCK2R may differ in each cancer as mutated CCKR2 has been detected in pancreatic cancer but not in colorectal or gastric cancers [58]. Further study is needed to explore CCKR2 as a therapeutic target in CRC.

Ghrelin

Ghrelin is an orexigenic neuropeptide, which apart from increasing appetite, it causes growth hormone (GH) release [59]. It is implicated in regulating metabolism, insulin release and inflammation [59]. Ghrelin and its receptor, Growth hormone secretagogue receptor 1α (GHSR1α) are barely detectable in normal colonic epithelial cells but are significantly higher in cancer cell lines (Caco2 and SW480 cells) and CRC tissue [59]. So far, the ghrelin-GHSR axis have not shown to have a prognostic role [60]. Inhibition of the GHSR1α using shRNA technology inhibits the growth of CRC cell line and xenografted tumour, suggesting GHSR1 can be a target for novel future treatment for early colorectal adenocarcinoma [59].

Guanylyl and uroguanylyl

Guanylyl and uroguanylyl are both peptide hormones secreted by the small and large bowel. They are involved in maintaining epithelial homeostasis and fluid secretion via the guanylyl cyclase (GCC2C) [61]. GCC2C are naturally found in colonic epithelial cells but not in other tissues [61]. GCC2C is expressed in 95% of primary and metastatic CRCs, whilst uroguanylin and guanylin expression disappears early on in CRC tumorigenesis [61].

Oral supplementation with uroguanylin has shown promise in preventing polyps in mice and CRC proliferation, without serious adverse effects [62]. A Phase I trial using Adenovirus 5-human guanylyl cyclase C-PADRE vaccine on 10 patients, showed that GCC2C vaccination is a safe and immunogenic target for cancer therapy [63]. Chimeric antigen receptors (CARs) used in mouse models with mestastatic CRC, eliminated metastases and increased survival, without toxicity [64]. Antibody-drug conjugates (ADCs), have also been used in a phase I clinical trial showing antitumor activity [65]. The commonest reported adverse effects included nausea, decreased appetite, fatigue, diarrhea, anaemia, alopecia and neutropenia [65]. Research into GCC2C target therapy is rapidly progressing and the next few years, will provide a better insight in its use in primary and metastatic CRC [61, 66].

Neurotensin (NT)

Neurotensin is a hormone mainly produced by the endocrine cells of the small bowel. Its main receptor, NTSR1, is undetectable in epithelial cells of normal colonic epithelium but moderately expressed in adenomas and adenocarcinomas [67]. Both in vivo and in vitro studies have shown that NT administration lead to colonic cancer proliferation [68].

Kamimae et al. [69] observed that epigenetic silencing of NTSR1 gene by methylation is associated with better prognosis in colorectal cancer. As the impact of methylation on a gene depends on which CpG islands have been studied, so further study is required to explore the prognostic potential in CRC [68].

SR 48692, a NT antagonist, inhibits the stimulatory effect of NT on colon cancer cells in xenografted mouse models and recently has shown promise as a potential therapeutic target in gastric cancer as well [68, 70]. Use of functionalized liposomes with NT peptide and filled with the cytotoxic doxorubicin in colon cancer cells, showed a better uptake of the drug and improved cytotoxicity [71]. Use of Neurotensin-Polyplex nanoparticles in murine cancer models has proven to be safe in patients with cancer expressing NTSR1 [72]. Clearly, NT has a role to play in CRC, but more work needs to be done to further explore its role in CRC.

Somatostatin/somatotropin-release inhibiting factor (SRIF)

Somatostatin is a peptide hormone produced by the GIT and the central nervous system and exists in 2 main forms: SRIF-14 and SRIF-28. Evangelou I [73] measured the expression of two of the somatostatin receptor subtypes (Sst1 and Sst5) in 81 patients, showing a negative association between CRC invasion and liver metastasis. Patients with either Sst2 or Sst5 expression had longer survival rates, although this was not an independent predictor of survival after controlling for other known prognostic factors [73]. Cytotoxic SRIF agonist, AN-238 has been shown to inhibit colon cancer growth in CRC cell lines and xenografts, both ofwhich express Sst, regardless of their p53 status [74]. Further study is however required to establish the role of somatostatin in CRC.

Neurokinin A, neurokinin B and substance P

Neurokinins (NK) and substance P (SP) are part of the tachykinin family, which are peptide hormones implicated in carcinogenesis. Chen et al. showed that high expression of its receptor, NK1R in CRC is associated with lymph node metastasis, TNM stage and worse survival [75]. Patients with higher SP expression also had a poorer prognosis (mean survival 43.53 ± 2.4 months) compared to patients with lower SP expression (mean survival time 79.98 ± 1.60 months) [75].

Garnier et al. [76] explored the therapeutic potential of NK1R antagonist, aprepitant (AP) by treating human colon cancer lines LiM6 (β catenin mutation) and DLD1 (APC mutation). Use of AP inhibited the growth of CRC cells through inhibition of the Wnt/β-catenin and AKT/mTOR signaling pathways. The NK1R antagonists also inhibited the cancer stem cells (CSCs), which are thought to confer higher resistance to chemotherapy and radiotherapy and therefore play a key role in CRC recurrence [76].

Central nervous system

Growth hormone

Growth hormone (GH) is a peptide hormone, secreted mainly by the pituitary gland. It acts directly on tissue or via IGF-1 through its action on the liver, promoting cell proliferation and differentiation [77]. Wang et al. [77] found that autocrine GH promoted cell proliferation, survival and oncogenicity of CRC cells in both in vitro and xenografted in vivo models. They also observed that the GH mRNA expression in CRC was associated with a larger tumor size and lymph node metastasis. Despite the evidence of GH implication in cancer, preclinical studies investigating the GHR antagonism as a therapeutic pathway for cancer treatment is limited.

Pegvisomant(B2036), which is an effective GHR blocker and used in patients with acromegaly, has shown limited use in preclinical studies because it behaves differently in non-human models [78]. Other alternative methods of inhibiting GHRs include: ATL1103, an antisense drug which blocks GHR (Antisense Therapeutics Ltd), a neutralizing GHR antibody and somatostatin agonists. In vitro and in vivo xenografted mouse model studies using siRNA knockdown of the GHR in human CRC cell line SW480, showed a reduced tumour growth and liver metastases [78]. Further studies, using GH antagonists should take into account, GH’s closely related sequence similarity to prolactin, its interaction with IGF1 receptors and epidermal growth factor receptors, to limit unwanted side-effects [78].

Conclusions

Colorectal cancer has shown to have a complex molecular pathway. The discovery of the role of hormones in breast and prostate cancer has changed the prognosis and treatment of its sufferers. Yet, hormonal links to CRC are not as clearly established. However, there is a lot of promise in the role of hormones in the prognostic and therapeutic role of hormones in CRC (Table 1). The link between obesity, T2DM and CRC is likely to be hormone-mediated in a certain subset of cancers. In addition, gut-derived hormones and their receptors show to be an attractive target, with uroguanylin/guanylin being the most promising. It is important to note that the action of hormones may share a common pathway in many of the other GI malignancies, posing a challenge in tailoring treatment strategies. Finally, this review has highlighted that only a small fraction of the hormones acting on the GIT has been investigated and those that have been investigated, are still mostly at an early stage of research and therefore need further study.

Table 1 Summary of hormones and their prognostic and therapeutic potential