Main

Heredity represents a major cause of colorectal cancer, with at least 20% of the cases estimated to develop because of genetic factors and ∼5% linked to inherited mutations in cancer-predisposing genes.1, 2, 3, 4 Worldwide, hereditary factors are annually estimated to cause 100 000 deaths from colorectal cancer. Patients whose tumors are identified at early stages have an excellent prognosis, and this implies an unprecedented possibility for disease prevention in individuals at increased risk. Regular colonoscopies have been demonstrated to effectively reduce morbidity and mortality from colorectal cancer in individuals with hereditary predisposition for the disease.5, 6

Hereditary colorectal cancer can broadly be divided into polyposis syndromes and nonpolyposis syndromes. The polyposis subset represents <1% of the cases and includes, for example, familial adenomatous polyposis (FAP) caused by ‘APC (adenomatous polyposis coli)’ mutations, MUTYH-associated polyposis (MAP) caused by ‘MUTYH (mutY homolog)’ mutations, Peutz–Jegher syndrome with ‘STK11 (serine/threonine kinase 11)’ mutations, and Juvenile Polyposis with mutations in ‘SMAD4 (SMAD family member 4)’ and ‘BMPR1A (bone morphogenetic protein receptor, type IA).’7, 8, 9, 10, 11, 12, 13 The term hereditary nonpolyposis colorectal cancer (HNPCC) was coined to distinguish familial aggregation of colorectal cancer from the polyposis phenotypes. The Amsterdam criteria (AC) were introduced for uniform classification based on family history and require at least three affected family members in two or more generations, with one being a first-degree relative of the other two and at least one individual diagnosed before 50 years of age.14, 15 The AC1 apply to families with ≥3 colorectal cancers and the AC2 also include extracolonic tumors, that is, endometrial cancer, cancer of the upper urinary tract, and cancer of the small bowel.14, 15 The AC are fulfilled in 2–5% of patients with colorectal cancer.4, 16, 17 The HNPCC subset of colorectal cancer families is heterogeneous and broadly consists of the 4% linked to Lynch syndrome (which may or may not fulfill the AC), <1% with a Lynch-like syndrome, and 2–4% classified as familial colorectal cancer type X (FCCTX). Lynch syndrome is defined by germline mismatch-repair (MMR) gene mutations in ‘MLH1 (mutL homolog 1),’ ‘MSH2 (mutS homolog 2),’ ‘MSH6 (mutS homolog 6), ‘PMS2 (PMS2 postmeiotic segregation increased 2, S. cerevisiae)’, but only about one-third of the Lynch syndrome families fulfill the AC criteria.18, 19 Lynch-like families are defined by the AC and show tumors with functional MMR gene defects (ie, loss of MMR protein expression and/or presence of microsatellite instability (MSI)), but lack of disease-predisposing MMR gene mutations.20 FCCTX families are defined as families that fulfill the AC1 and show MMR-stable tumors and lack of MMR gene mutations.4, 16, 19, 21 Some studies of FCCTX have also included AC2 families with MMR-stable tumors.22, 23, 24, 25 The FCCTX subset is a major cause of hereditary colorectal cancer, although it remains a weakly defined and sparsely investigated subgroup of hereditary colorectal cancer. A better understanding of hereditary colorectal cancer may provide important clues to disease-predisposition and could contribute to molecular diagnostics, improved risk stratification, and targeted therapeutic strategies. These needs motivate our review of the clinical presentation, histopathologic features, and molecular mechanisms of FCCTX.

Clinical presentation, histopathologic features, genetic susceptibility, and recommendations for surveillance

Registry data from Denmark, Finland, Italy, Australia, and from the Epicolon study suggest that 21–73% of the families that fulfil the AC1 and/or AC2 represent FCCTX16, 26, 27, 28 (Figure 1). Age at onset shows considerable interfamily as well as intrafamily variability with colorectal cancer diagnosed at a higher mean age (mean 57.3 years) in FCCTX than in Lynch syndrome (mean 49.7 years; Figure 1).16, 22, 29, 30, 31, 32 The FCCTX tumor spectrum is predominated by colorectal cancer and does, in contrast to Lynch syndrome, not show any increased risk of extracolonic cancers.16, 33, 34 Colorectal cancers linked to FCCTX are left sided in 70% of the cases. Synchronous as well as metachronous adenomas are frequent with a high adenoma/carcinoma ratio that may suggest a slower adenomacarcinoma progression rate than in Lynch syndrome.16, 21, 31, 33, 35, 36, 37, 38, 39

Figure 1
figure 1

(a) Frequency of Amsterdam I/II families without disease-causing germline MMR mutations among different registers. (b) Mean age at diagnosis in Lynch Syndrome and FCCTX among the different registers. Source: Danish Hereditary NonPolyposis Register (Denmark), CCFR: Colorectal Cancer Family Register (Australia), Finnish Hereditary Register (Finland), Epicolon Study and Register of Hereditary Digestive Tumors at Foundation IRCCs National Cancer Institute of Milan (Italy).

Full size image

Whereas colorectal cancers linked to Lynch syndrome are characterized by poorly differentiated tumors, mucinous differentiation, an expanding growth pattern, and abundant lymphocytic reactions (including tumor-infiltrating lymphocytes, peritumoral lymphocytes, and Crohn-like reactions), FCCTX tumors typically show a more ‘sporadic-like’ phenotype with medium high differentiation, glandular and infiltrative growth patterns, and frequent dirty necrosis (Figure 2).37,38 The lack of distinct histopathologic features makes identification of FCCTX-associated colorectal cancer challenging from a pathologist’s perspective and underscores the importance of obtaining a family history of cancer.

Figure 2
figure 2

Hematoxylin and eosin-stained slides from colorectal cancers linked to FCCTX demonstrating (a) dirty necrosis and (b) infiltrative growth in the deep tumor margin. Dirty necrosis is characterized by the presence of large amounts of cell detritus and inflammatory cells within the glandular lumina, whereas the infiltrative growth pattern is characterized by the widespread dissemination of intraepithelial, mainly cytotoxic, T lymphocytes within the tumor tissue. Photo is courtesy of Professor Susanne Holck, Hvidovre Hospital, Denmark.

Full size image

Surveillance programs in FCCTX are targeted at colorectal cancer and the mean age at onset of 60 years implies that surveillance colonoscopies are generally recommended with 3–5-year intervals, starting 5–10 years before the earliest age at onset in the family.3, 16, 19, 34, 40

A number of genome-wide association studies have addressed susceptibility loci in hereditary colorectal cancer, although not specifically linked to FCCTX. Candidate genetic variants have been reported in ‘CENPE (centromere protein E, 312 kDa)’ at 4q24–q25, ‘CDH18 (cadherin 18, type 2)’ at 5p14.3, ‘GALNT12 (UDP-N-acetyl-α-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase 12, GalNAc-T12),’ ‘ZNF367 (zinc finger protein 367),’ ‘HABP4 (hyaluronan binding protein 4),’ and ‘GABBR2 (γ-aminobutyric acid (GABA) B receptor, 2)’ on chromosome 9, and ‘BMP4 (bone morphogenetic protein 4)’ at 8q23.3, ‘GREM1 (gremlin 1, DAN family BMP antagonist)’ and ‘KIF24 (kinesin family member 24)’ on chromosome 15, and ‘BCR (breakpoint cluster region)’ at 22q11.23, 41, 42, 43, 44, 45, 46 In addition, two microRNA genes, hsa-mir-491/KIAA1797 and hsa-mir-646/AK309218, have been associated, but additional data and independent validations are needed for application of these markers in clinical risk models.46 Current understanding suggests that the risk alleles identified are insufficient to independently account for FCCTX, but a combination of moderate and low-risk alleles could contribute to the familial aggregation.4, 39, 40, 41, 42, 47

Genomic and epigenetic profiles of FCCTX-associated colorectal cancer

Deranged DNA methylation is inversely associated with the MSI and the CpG island methylation phenotypes (CIMP) and has been demonstrated in 30–40% of colorectal cancers.48, 49, 50, 51, 52 Hypomethylation in long interspersed nucleotide element-1 (LINE-1) has been linked to familial CRC, including FCCTX, and is thought to interfere with chromosomal segregation and thereby enhance chromosomal instability (CIN).49, 51, 53, 54, 55, 56 Global DNA hypomethylation has been associated with poor prognosis, shorter survival, younger age of onset, and familial colorectal cancer risk.39, 51, 55, 56, 57, 58, 59 The predisposition to LINE-1 hypomethylation in FCCTX tumors gives an evidence for a link between distinct molecular signatures and phenotypes associated with specific epigenotypes.59 In hereditary colorectal cancer, relatively less is known about the patterns of specific histone modifications that also regulate gene expression through controlling chromatin conformation.52, 60 Recent studies have shown that mutation rates in colorectal cancer genomes are closely related to histone modification-directed chromatin organization and its upregulation is associated with a reduced patient survival,52, 61 suggesting a key role in colorectal cancer development.

The genomic profiles of FCCTX tumors show similarities to sporadic MMR-proficient colorectal cancer.42 Comparative genomic hybridization studies suggest that FCCTX tumors typically harbor 6–8 copy number alterations with recurrent gains of 7p, 7q, 8q, 13q, 20p, and 20q and losses of 17p, 18p, and 18q.62, 63 Gain of the 20q region has been specifically linked to FCCTX tumors and several candidate target genes such as ‘GNAS (GNAS complex locus),’ ‘AURKA (aurora kinase A),’ ‘SRC (v-src avian sarcoma (Schmidt-Ruppin A-2) viral oncogene homolog),’ ‘TOP1 (topoisomerase (DNA) I),’ ‘NELFCD (negative elongation factor complex member C/D),’ ‘ADRM1 (adhesion regulating molecule 1),’ ‘ASIP (agouti signaling protein),’ ‘CDH26 (cadherin 26),’ and ‘HNF1A (HNF1 homeobox A)’ reside herein (Figure 3).24, 62, 63 GNAS promotes proliferation through activation of the Wnt and ERK1/2 MAPK signaling pathways.64 Activating missense mutations in GNAS have been demonstrated in sporadic colorectal cancer, but the GNAS c.601G>T hot spot mutation could not be identified in a study of FCCTX tumors.63, 65 An overexpression of the chromosome-associated gene AURKA has been correlated to aneuploidy, invasion, and progression of colorectal adenoma to carcinoma.66 A candidate link also exists between gain of 20q and overexpression of ASIP, that could influence sensitivity to 5-fluorouracil.67 Loss of chromosome 18 is a common change in sporadic tumors as well as in FCCTX tumors and may be linked to downregulation of for example, ‘SMAD2 (SMAD family member 2),’ ‘SMAD4 (SMAD family member 4),’ ‘DCC (deleted in colorectal carcinoma),’ ‘SERPINB5 (serpin peptidase inhibitor, clade B (ovalbumin), member 5),’ and ‘BCL2 (B-cell CLL/lymphoma 2)’ (Figure 3).24, 25 Frequent (32%) genome-wide copy neutral loss of heterozygosity and a low frequency (14%) of chromosomal losses have been demonstrated in FCCTX tumors that could indicate involvement of yet unidentified DNA repair mechanisms.63, 68

Figure 3
figure 3

Model illustrating identified genetic alterations that may promote tumorigenesis in FCCTX tumors. FCCTX are chromosomally instable with gains in chromosomal region 20q and complete loss of chromosome 18. These alterations may result in deregulation of genes involved in chromosomal segregation and genomic instability (AURKA), apoptosis (BCL2 and SERPINB5), proliferation (GNAS), growth inhibition, angiogenesis (PTGER1), and migration (SRC).

Full size image

Mutations in several cancer-related genes such as ‘TP53 (tumor protein p53),’ ‘KRAS (Kirsten rat sarcoma viral oncogene homolog),’ ‘BRAF (v-raf murine sarcoma viral oncogene homolog B),’ APC, ‘MGMT (O-6-methylguanine-DNA methyltransferase),’ and ‘CTNNB1 (catenin (cadherin-associated protein), β1, 88 kDa)’ divide FCCTX tumors into two major groups; one-third of the tumors that are characterized by stable genotypes with few genetic changes retained membranous β-catenin expression and infrequent TP53 mutations, and two-thirds of the tumors with frequent loss of tumor suppressor gene loci such as APC, TP53, SMAD4, and DCC, somatic methylation of APC, KRAS, and MGMT, and nuclear translocation of β-catenin.22, 69 These genetic subsets have been suggested to differ in clinical presentation; genetically simple tumors predominantly develop in the proximal colon and develop at a lower (mean 54 years) age, whereas the genetically complex tumors more often develop in the distal colon and are diagnosed at a higher (mean 59 years) age.

Gene expression profiles and deranged signaling pathways

Gene expression studies in colorectal cancer have predominantly focused on differences between sporadic MMR-proficient and MMR-deficient tumors, whereas data on FCCTX tumors are scarce.70 MMR-proficient and MMR-deficient tumors show distinct profiles with 65–2070 significantly deregulated genes, including genes involved in growth factor receptors, transcription, cell cycle function, DNA repair, chromatin structure, drug metabolism, and chemoresistance.70, 71, 72, 73, 74, 75, 76 Gene expression data from FCCTX tumors suggest similarity to sporadic MMR-proficient colorectal cancers with upregulation of genes involved in peptidyl-amino acid modification, enzyme-linked receptor protein signaling, growth regulation, DNA repair pathways, vascular smooth muscle contraction, and G protein-coupled receptor signaling.25, 77 The limited data available regarding signaling pathways in FCCTX tumors indicate involvement of G protein-coupled signaling and candidate genes involved in proliferation and migration, for example, CDH26, SRC, and ASIP (located in chromosome 20q).24, 25 ‘PTGER1 (prostaglandin E receptor 1 (subtype EP1), 42 kDa)’ is activated by COX-2/PGE-2 signaling that is overexpressed in 90% of sporadic colon carcinomas and is induced by hypoxia.78, 79 Another target that has shown to directly upregulate COX-2 expression is the ‘HIST1H1A (histone cluster 1, H1a)’, whose overexpression had been significantly associated with a shorter colorectal cancer-specific and overall survival.80 Through a SRC-dependent pathway, PTGER1 activates the ‘ANGPTL4 (angiopoietin-like 4)’ protein that may promote colorectal carcinogenesis.79 Upregulation of this pathway may drive FCCTX tumor development during anaerobic conditions, and this is supported by frequent findings of dirty necrosis during histological evaluations and downregulation of the aerobic oxidative phosphorylation metabolism genes such as ‘ATP5L (ATP synthase, H+ transporting, mitochondrial Fo complex, subunit G),’ ‘ATP5A1 (ATP synthase, H+ transporting, mitochondrial F1 complex, α-subunit 1, cardiac muscle),’ ‘ATP5B (ATP synthase, H+ transporting, mitochondrial F1 complex, β-polypeptide),’ and ‘ATP5D (ATP synthase, H+ transporting, mitochondrial F1 complex, δ-subunit)’ (Figure 3).25, 37 Copy number gains and EGFR-mediated activation of SRC have been correlated to migration and invasion81 (Figure 3). The TGFβR-mediated growth inhibition is executed through SMAD2 and SMAD4, whereas BCL2 induces apoptosis. Hence, biallelic loss of these genes may give FCCTX tumors a growth advantage (Figure 3). The tumor suppressor ‘SERPINB5 (serpin peptidase inhibitor, clade B (ovalbumin), member 5)’ is activated by TP53 and induces apoptosis and inhibits migration and invasion of tumor cells.43 The current understanding of how gene expression profiles in FCCTX-associated tumors influence tumor development thus suggests increased proliferation, reduced apoptotic activity, and enhanced migration and invasion that could contribute to the poor prognosis suggested in this subgroup (Figure 3).

Conclusion

The FCCTX subset is challenging, not least as the clinical presentation and the histopathologic features mimic sporadic MMR-proficient tumors. The genetic causes are unknown, but candidate genes include, for example, CENPE, CDH18, GREM1, BCR, KIF24, GALNT12, ZNF367, HABP4, GABBR2, and BMP4. Differences in genomic and gene expression profiles do exist, for example gain of chromosome 20q, global hypomethylation, and upregulation of the G-protein coupled receptor signaling pathway. Taken together, current data suggest tumor development linked to inhibition of apoptosis, insensitivity to growth inhibitory signals, inhibition of angiogenesis, and increased migration and invasion that may be reflected in infiltrative growth patterns and frequent dirty necrosis. Extended and in-depth analyses of the FCCTX tumor genome, methylome, and proteome could in well-defined tumor series shed light on the basic mechanisms for potential application in refined diagnosis and targeted interventions.