61. Молекулярная биология колоректального рака


Колоректальный рак (CRC) является прототипом идеи о том, что солидные опухоли возникают и прогрессируют в результате дефектов ДНК репарации или стабильности хромосом и аккумуляции соматических мутаций и эпигенетических изменений. Это один из немногих типов рака, который развиваются у животных из специфической нативной популяции интерстициальных стволовых клеток (ISC). Эта двойная оценка молекулярных дефектов и клетки происхождения глубоко обоснована исследованиями общих соматических мутаций, хорошо охарактеризованных синдромов предрасположенности и сигналов, которые поддерживают нормальные ISC.

Многоступенчатые модели колоректального рака и генетической нестабильности

Nearly all CRCs develop within benign precursor polyps, where gatekeeper mutations initiate epithelial overgrowth by constitutive activation of the Wnt signaling pathway and additional mutations combine to promote invasion and metastasis (Fig. 61.1A). Pedunculated polyps larger than 1 cm confer the highest risk, with approximately 15% progressing to invasive cancer over 10 years. The prevalence of polyps in the U.S. population, estimated at 50% by age 70 years, dwarfs the 5% lifetime risk of developing CRC because few polyps become invasive and aberrations that promote malignancy accumulate over decades.1 Nevertheless, endoscopic resection of adenomas reduces CRC incidence and mortality. Hyperplastic polyps confer low cancer risk, but approximately 8% of sporadic CRCs originate in sessile adenomas with serrated histology, characteristic nuclear morphology, and little to no dysplasia.2 These CRCs may represent a distinct subtype, with high microsatellite instability (MSI- hi), frequent BRAF mutation, the CpG island methylator phenotype (CIMP), and a poor prognosis (Fig. 61.1B).

Inherited CRC syndromes and sporadic tumors together highlight the importance of proper DNA repair in preventing CRC. Scores to thousands of somatic DNA aberrations accumulate as a result of hallmark genetic instability, which is acquired in a limited number of stereotypic ways (see Fig. 61.1). About 80% of CRCs show widespread chromosomal instability (CIN): gains, losses, and translocations that produce various gene amplifications, deletions, and rearrangements. Chromosomal segregation defects, mediated by segregation factors such as BUB1, may underlie CIN, but few genes are implicated directly. CRCs with CIN show a genome-wide bias toward C:G to T:A transitions at 5′-CpG-3′, and fewer mutations in 5′-TpC-3′, dinucleotides than breast cancer, for example.3 On average, 17 genes are deleted or amplified to 12 or more copies per CRC,4 with deletions resulting in loss of heterozygosity (LOH). In aggregate, the oncogenes ERBB2, MYC, KRAS, MYB, IGF2, CCND1, and CDK8 are amplified or overexpressed in most cases, usually along with neighboring genes, but nearly half the copy number alterations also occur in other cancers. CRCs thus reflect perturbation of selected pathways of replicative and tissue homeostasis, some common to many cancers and others restricted to CRC. Specific cytogenetic anomalies correlate poorly with clinical outcomes or disease features. Recent evidence suggests that beyond driving widespread gene alterations, CIN causes excessive spillage of DNA into the cytosol, hence activating noncanonical nuclear factor kappa B (NF-κB) signaling, which promotes metastasis.5

Фигура 61.1. Генетические пути колоректальной карциномы. All colorectal cancers (CRCs) arise within benign adenomatous precursors, fueled by mutations that serially enhance malignant behavior. Mutations that activate the Wnt signaling pathway seem to be necessary initiating events, after which two possible courses contribute to the accumulation of additional mutations. A:  Chromosomal instability is a feature of up to 80% of CRCs and is commonly associated with activating KRAS point mutations and loss of regions that encompass P53 and other tumor suppressors on 18q and 17p, often but not necessarily in that order. B: About 20% of CRCs are euploid but defective in DNA mismatch repair (MMR), resulting in high microsatellite instability (MSI-hi). MMR defects may develop sporadically, associated with CpG island methylation (CIMP), or as a result of familial predisposition in hereditary nonpolyposis colorectal cancer (HNPCC).

Mutations accumulate in the KRAS or BRAF oncogenes, in p53 tumor suppressor, and in microsatellite-containing genes vulnerable to MMR defects, such as TGFβIIR. Epigenetic inactivation of the MMR gene MLH1 and activating BRAF point mutations are especially common in serrated adenomas, which progress, in part, through the silencing of tumor suppressor genes by promoter hypermethylation. Progression from adenoma to CRC takes years to decades, a process that accelerates in the presence of MMR defects. CIN, chromosomal instability.

Although the remaining approximately 15% of CRCs are euploid, they carry thousands of small insertions and deletions (indels) or point mutations near nucleotide repeat tracts—collectively designated as MSI-hi—as a result of defective DNA mismatch repair (MMR; see Fig. 61.1B).6 Up to one-third of these cases occur in the setting of the familial Lynch syndrome, discussed at length in the following text. MSI-hi tumors usually arise in the ascending colon, have a high frequency of BRAFV600E mutation, resist adjuvant 5-fluorouracil treatment, and respond better to immunotherapy with programmed cell death protein 1 (PD-1) blockade than other cancers.7 Stage for stage, aneuploidy in CIN confers a worse prognosis than MSI-hi disease.8 Some authors consider CIMP- positive serrated CRCs separately, but because their molecular, clinical, and pathologic features overlap with MSI-hi tumors, The Cancer Genome Atlas (TCGA) and other groups pool the two CRC types into a single “hypermutated” category.9

Among the approximately 3% of sporadic CRCs cases that are hypermutated but lack MSI, the majority carry somatic mutations in the exonuclease domain of POLE.9 Moreover, rare kindreds with a dominantly inherited high risk of CRC carry specific germline defects in POLE or POLD1, the proofreading exonuclease genes for the leading- and lagging-strand DNA polymerases ε and δ. Although the tumors carry thousands of mutations, microsatellite tracts are stable; polymerase proofreading-associated polyposis (PPAP) thus represents a third “mutator” mechanism in CRC.

Мутационные и эпигенетические ландшафты в колоректальном раке

After accounting for the different paths to genome instability, mutational profiles are similar in colon and rectal cancers. Constitutive activation of the Wnt pathway—triggered by gatekeeper mutations in APC, CTNNB1, RNF43, or RSPO genes—initiates adenoma formation, as discussed in the following text. Although the classic progression sequence (see Fig. 61.1) has its origins in the frequencies of various mutations at different stages of disease, CRCs vary substantially in which genes are mutated and in which order. Early studies estimated an average of 81 mutations per CRC, with a few mutations (APC, KRAS, TP53, PIK3CA) occurring frequently, others of known function (e.g., BRAFV600E) less commonly, and most mutations detected only in a handful of cases.3 Genome-scale studies of hundreds of CRCs and matched normal colonic mucosa corroborate these findings and reliably catalog the genetic landscape of CRC (Table 61.1).9,10 Few novel mutations uncovered in the latter studies lie in genes that constitute “druggable” targets, such as kinases, and some may have pleiotropic effects on cell survival, growth, and metastasis. In aggregate, genes that modify the epigenome represent the largest new class of frequently mutated genes, and MSI-hi tumors in particular show recurrent mutations in ARID1A, a chromatin remodeling factor. CRCs arising in African Americans have a partially distinct profile, including mutations in EPHA6 and FLCN,11 and approximately 10% of CRCs carry monoallelic missense mutations in SOX9, a transcription factor that is highly expressed and active in intestinal crypt stem and progenitor cells.17

CRCs progress by activating or disrupting pathways that involve many gene products; some genes in vital pathways are more prone to mutation than others, and mutation of a crucial gene may obviate the need for additional mutations in the same pathway. Thus, KRAS and BRAF mutations, which together occur in about half of all CRCs but are mutually exclusive, represent alternative means to activate intracellular epidermal growth factor receptor (EGFR) signaling. Although these examples conform to the paradigm that frequent occurrence of a genetic aberration reflects the selective advantage it confers on tumor cells, the mere presence of other mutations does not necessarily signify a pathogenic role. “Passenger” alterations, whether associated with CIN or MSI, confer no advantage and may even be detrimental to tumors. Two features are therefore used to impute “driver” status: recurrent alteration, as detected in large cohorts, and experimental demonstration of a contribution to malignancy, which is laborious and imperfect. Infrequent events that may contribute to neoplasia tend to concentrate in genes that control cell adhesion, signaling, DNA topology, and the cell cycle. Common mutations in sporadic CRCs rarely correlate with specific histologic or clinical features, but certain genotypes do delineate disease subtypes and sensitivity to certain therapies. For example, KRAS and BRAF mutations preclude clinical response to EGFR antibodies12,13 and MSI-hi CRCs are uniquely sensitive to treatment with immune checkpoint inhibitors.14

Prominent modes of epigenetic control in mammalian cells include DNA methylation at CpG dinucleotides and covalent modification of certain histone residues. Compared to normal colonic epithelium, CRCs and even benign adenomas show 8% to 15% lower total DNA methylation.15,16 Global DNA hypomethylation may reduce the fidelity of chromosome segregation, owing to reduced pericentromeric methylation, and reversal of imprinting at loci such as IGF2, but its pathogenic role, if any, is unclear. In mouse models, some studies suggest that global hypomethylation increases tumor susceptibility, whereas absence of the de novo methyltransferase DNMT3B slows, and DNMT3B overexpression accelerates, tumor progression. Against the backdrop of genome-wide DNA hypomethylation, the distinctive subset of CRCs with CIMP shows coordinate hypermethylation of many CpG island promoters, associated with transcriptional attenuation of tumor suppressor genes such as HIC1 and Wnt- inhibiting secreted frizzled related protein (SFRPs).17,18 Whole-genome methylation analyses affirm the existence of this once controversial entity,19 revealing properties that distinguish it from KRAS-mutant CIN-positive disease: origin in sessile serrated adenomas; strong association with BRAF-mutant, right-sided MSI-hi tumors with MLH1 gene methylation; and distinct RNA expression profiles.9

Таблица 61.1. Рекуррентные мутации соматических генов в колоректальном раке человека

Ген Частота CRC класс Известная клеточная функция
KRAS 35%–40% CIN RTK сигналинг
PIK3CA 18%–20% Mostly CIN RTK сигналинг
BRAF 7%–15% MSI-hi, CIMP RTK сигналинг
NRAS 9% CIN RTK сигналинг
ERBB3 <8% CIN RTK сигналинг
CTNNB1 <5% All Активация Wnt пути
Опухоль-супрессорные гены
APC 85% All Регуляция Wnt пути
RNF43 18% Mainly MSI-hi Регуляция Wnt пути
TP53 50% Mainly CIN Stress, hypoxia response
SMAD4 10% CIN TGF-β сигналинг
FBXW7 10% CIN Ubiquitin ligase
SOX9 5%–10% CIN Wnt-зависимая функция ISC
TGFBR2 MSI-hi TGF-β сигналинг
MSH2, etc. MSI-hi DNA mismatch repair
POLE Hypermutated (non-MSI) DNA polymerase ε
Эпигенетически модифицируемые гены (роли проясняются)
ARID1A MSI-hi Перестройка хроматина
SIN3A Репрессия транскрипции
SMARCA5 Перестройка хроматина
NCOR1 Репрессия транскрипции
JARID2 Модификация гистонов
TET1, 2, 3 ДНК деметилирование

CRC, colorectal cancer; CIN, chromosomal instability; RTK, receptor tyrosine kinase; MSI-hi, high microsatellite instability; CIMP, CpG island methylator phenotype; TGF-β, transforming growth factor β; ISC, intestinal stem cell.

Understanding of other epigenetic alterations in CRC is still in its infancy. Differences in the landscape of modified histone residues from that in normal colonic epithelium may reflect epigenetic rewiring, including aberrant activation of cis-elements, or the different proportions of stem and progenitor cells in tumors and normal tissue. The product of one gene, ARID1A, which is commonly mutated in MSI-hi CRCs, normally targets the SWI/SNF chromatin remodeling complex to enhancer elements. In ARID1A-deficient CRC cell lines and mouse colonic epithelium, defective SWI/SNF targeting causes widespread enhancer and gene dysregulation.20

Понимание кишечной крипты мыши и колоректального рака человека приводят к модели последовательной инициации и прогрессирования колоректального рака

WNT сигналинг

ISCs in the human colon and mouse small intestine have a fundamentally similar physiology, including dependence on Wnt signaling, the pathway that is confidently implicated in initiating CRC. One informative strain, multiple intestinal neoplasia (Min), carries a truncating mutation in Apc and phenocopies human familial adenomatous polyposis (FAP),21 although adenomas form mainly in the small intestine and not in the colon. Deletion of the N-terminal domain of mouse β-catenin also mimics Wnt signaling and induces intestinal polyposis. ApcMin and other mouse strains are thus invaluable for studying and modeling CRC.

Continuous epithelial renewal in mouse intestinal crypts relies on frequent replication of native LGR5-expressing ISC, with neutral drift producing stable populations of six to eight monoclonal stem cells in each crypt. Although crypt progenitors are strikingly able to dedifferentiate into ISC when native stem cells are injured, LGR5-positive ISC are uniquely vulnerable to transformation by Wnt pathway activation,22 and LGR5-positive cells display stem cell properties in CRC xenografts.23 Thus, most human and murine CRCs likely arise from cells at the top of the lineage hierarchy and not from their descendants. Selective depletion of LGR5- positive tumor cells, however, causes their LGR5-negative progeny to assume long-term self-renewal,24 revealing a remarkably fluid cancer stem cell compartment that mirrors normal ISC plasticity. Ligand-independent ISC with mutations that activate Wnt signaling compete by neutral drift with their wild-type siblings in the same crypt, and in this contest, the mutations endow ISC with a small, measurable growth advantage.25 Once a mutant ISC clone takes hold, free from competing wild-type ISCs in the same crypt, it may flourish indefinitely and accumulate additional mutations, by chance or by virtue of hypermutable states. Over time, various combinations of mutations liberate signaling circuits from other ligand dependencies and impart invasive properties, an outcome that is not inevitable: Few polyps progress to carcinoma, and others may regress.

Wnt glycoproteins have diverse developmental and homeostatic functions. Secreted after palmitoylation by the acyltransferase PORCN, their intestinal activity is markedly potentiated by R-spondins (RSPO), which are ligands for the ISC marker and surface receptor LGR5.26,27 In the absence of Wnt stimulation, an intracellular complex containing APC, AXIN2, and other proteins enables phosphorylation of conserved N-terminal serine and threonine residues in CTNNB1 (β-catenin). This phosphorylation targets a cytosolic pool of β-catenin, distinct from the abundant stores bound to E-cadherin on the inner cell membrane, for ubiquitin-mediated proteasomal degradation (Fig. 61.2). Binding of Wnts to a surface complex containing a FRIZZLED (FRZD) protein and the obligate coreceptor LRP5/6 inhibits the APC-AXIN2 destruction complex. Thus stabilized, β-catenin shifts to the cell nucleus, where it functions as an obligate coactivator for target genes of the TCF family of transcription factors, especially TCF7L2. Important and pertinent negative regulators of Wnt signaling include two related transmembrane E3 ubiquitin ligases, RNF43 and ZNRF3, which induce endocytosis and degradation of FRZD receptors.28 The LGR5/RSPO complex potentiates Wnt signaling by neutralizing these ligases, hence increasing FRZD receptor availability.29 This detailed elucidation of Wnt pathway circuitry explains why CRCs require either inactivating mutations in the tumor suppressor genes APC or RNF43 or events that activate the CTNNB1, RSPO2, RSPO3, or, rarely, TCF7L2 oncogenes.

Фигура 61.2. Схема Wnt сигнального пути. Липид-модифицированные Wnt гликопротеины связывают Frizzled-LRP5/6 корецепторный комплекс для ингибирования декструктивного комплекса, который включает AXIN и продукт APC (Adenomatous Polyposis Coli) гена. This complex ordinarily targets β-catenin for proteasomal degradation, and its inhibition in response to Wnt signaling stabilizes cytosolic β-catenin, which translocates to the nucleus and coactivates transcriptional targets of the T-cell factor/lymphoid enhancer factor (TCF/LEF) family of sequence-specific transcription factors. The net result in intestinal crypt cells, mediated in part by target genes such as MYC, is to foster cell replication. Important regulatory inputs to this signaling pathway are provided R-spondins (RSPO), potent hormonal potentiators of Wnt signaling. RSPO ligands bind to LGR5/6 surface receptors and inhibit the transmembrane ubiquitin ligase RNF43, which otherwise degrades Frizzled molecules.

Through inhibition of RNF43, the net result of RSPO signaling is to increase Frizzled receptor density on the cell surface, increasing cell sensitivity to ambient Wnt concentrations. Mutations found in human colorectal cancers affirm that APC and RNF43 are tumor suppressor genes (red balloons), whereas RSPO, β-catenin (CTNNB1), and the transcription factor TCF7L2 behave as oncogenes (green balloons).

Другие пути ростовых факторов

Beyond Wnt signaling, normal colon crypts rely on a correct balance of growth signals, principally from the epidermal growth factor (EGF) family, and antimitotic signals from the bone morphogenetic protein (BMP) family of transforming growth factor β (TGF-β) ligands. Genes commonly mutated in sporadic CRC (KRAS, PIK3CA, or BRAF) encode intracellular mediators of EGF signaling. The resulting ligand-independent mode of constitutive intracellular pathway activation differs fundamentally from that of EGFR point mutations in lung cancer, and reflecting this key difference, CRCs that lack KRAS or BRAF mutations respond to anti-EGFR antibodies, but not to small-molecule EGFR inhibitors such as gefitinib and erlotinib; EGFR-mutant lung cancers show the converse sensitivities.

BMP ligands signal through specific cell surface receptors to activate the SMAD family of latent transcription factors, and the tumor suppressor gene mutations in some familial predisposition syndromes concentrate in this signaling pathway. Loss of BMP function in mice expands stem and progenitor cells, leading to polyposis or ectopic crypts,30–32 and cultured CRC cells resist antiproliferative effects of TGF-β.33 Enteroid cultures of normal human and mouse colonic epithelium require Wnt/R-spondin, EGF, and BMP antagonists, and sequential introduction of specific APC, KRAS, and SMAD mutations eliminates the respective requirements.24,34–36 Recurrent mutations thus reveal that CRC coopts pathway dependencies within the normal tissue’s signaling circuits; because normal colonic and other cells rely on the same pathways, targeting them in the clinic has inherent limits.

Наследственные синдромы высокого риска рака проясняют ранние реакции и критические пути в колоректальном онкогенезе

Two Mendelian syndromes, FAP and hereditary nonpolyposis CRC (HNPCC), together account for approximately 5% of CRCs; other syndromes, each occurring in fewer than 1 in 200,000 births, also elevate the risk (Table 61.2). Moreover, up to one-quarter of cases may have an unappreciated familial basis. Beyond improving molecular diagnosis, risk assessment, and targeted prevention in affected kindreds, knowledge about predisposing conditions profoundly informs understanding of the much larger fraction of sporadic cases.

Семейный аденоматозный полипоз и центральное значение WNT сигналинга

FAP, an autosomal dominant monogenic disorder, underlies approximately 0.5% of all CRCs. Individuals develop hundreds to thousands of colonic polyps by their early 20s, with 100% lifetime risk of CRC. Benign extraintestinal manifestations include duodenal and gastric adenomas; congenital hypertrophy of the retinal pigmented epithelium; osteomas and mesenteric desmoid tumors in the Gardner syndrome variant; brain tumors in the Turcot syndrome; and, occasionally, cutaneous cysts, thyroid tumors, or adrenal adenomas. The responsible gene, adenomatous polyposis coli (APC), encodes a 300-kDa component of the β-catenin destruction complex in the Wnt pathway. Truncating germline mutations tend to cluster in the 5′ half and exon 15. A few mutations correlate with phenotypic severity or specific extraintestinal features, but identical mutations can produce distinct clinical manifestations. Mutations in the extreme 5′ or 3′ ends of APC exons cause a variant syndrome, attenuated APC, in which few polyps or CRCs develop late in life. Identification of specific APC mutations in probands allows reliable testing of family members. Minimal recommendations for carriers include annual screening colonoscopy after age 10 years and treatment with nonsteroidal anti-inflammatory drugs to reduce CRC risk. Colectomy is highly recommended for CRC prophylaxis, with continued lifelong vigilance over the rectal stump and other at-risk tissues. Reflecting the aforementioned similarities in crypt homeostasis in the small intestine and colon, patients have a 5% to 10% risk of developing periampullary adenocarcinoma, which mandates surveillance gastroduodenoscopy after age 25 years.

Таблица 61.2. Синдромы наследственного колоректального рака

Синдром Типично наблюдаемые признаки у больных Вовлеченный ген
Синдромы с аденоматозными полипами
FAP Multiple adenomas (>100) and colorectal carcinomas; duodenal polyps and carcinomas; gastric fundus polyps; congenital hypertrophy of retinal epithelium APC (>90%)
Gardner syndrome Same as FAP, with desmoid tumors and mandibular osteomas APC
Turcot syndrome Polyposis and CRC with brain tumors (medulloblastoma, glioblastoma) APC or MLH1
Attenuated FAP Less than 100 polyps, although marked variation in polyp number (from <5 to >1,000 polyps) seen in mutation carriers within a single family APC (5′ мутации)
HNPCC CRC with modest polyposis; high risk of endometrial cancer; some risk of ovarian, gastric, urothelial, hepatobiliary, and brain cancers MSH2, MLH1, MSH6 (вместе >90%); PMS2 (около 5%)
MAP Multiple gastrointestinal polyps, autosomal recessive MYH
Polymerase proof reading-associated polyposis Large adenomas, early-onset CRC, elevated risk of endometrial cancer only POLE или POLD1
Синдромы с атипичными полипами
Peutz-Jeghers syndrome Hamartomatous polyps throughout the GI tract; mucocutaneous pigmentation; estimated 9–13-fold increased risk of GI and non-GI cancers STK11 (30%–70%)
Cowden disease Multiple hamartomas involving breast, thyroid, skin, brain, and GI tract; increased risk of breast, uterus, thyroid, and some GI cancers PTEN (85%)
Juvenile polyposis syndrome Multiple hamartomas in youth, predominantly in colon and stomach; variable increase in colorectal and stomach cancer risk; facial changes BMPR1A (25%), SMAD4 (15%), ENG
Hereditary mixed polyposis Polyps of highly heterogeneous form and size, a few of which progress to CRC; confined to rare Ashkenazi Jewish kindreds; only CRC risk is elevated GREM1 (imputed)


FAP — семейный аденоматозный полипоз; CRC — колоректальный рак; HNPCC — наследственный неполипозный колоректальный рак; MAP — MYH-ассоциированный полипоз; GI — гастроинтестинальный.

The larger significance of APC stems from its somatic inactivation in approximately 80% of sporadic CRCs. Both familial and sporadic CRCs show biallelic inactivation, with one copy usually lost by deletion. It is the earliest genetic aberration in adenomas,37 including tiny polyps with minimal dysplasia, and the rate-limiting step in polyp formation; indeed, acute APC loss in mice immediately activates the Wnt pathway, resulting in epithelial hyperproliferation.38 As APC encodes many functional domains, mutant forms might affect diverse cellular activities, contributing to chromosome segregation defects and aneuploidy. However, about half the sporadic CRCs with intact APC function have activating point mutations in CTNNB1,39,40 and many of the rest (approximately 10% of cases) carry gene fusions involving R-spondin (RSPO) cofactors.10 Therefore, attention on APC centers appropriately on its role in Wnt pathway inhibition. CTNNB1 mutations in CRC target residues for N-terminal phosphorylation, rendering β-catenin resistant to degradation, and RSPO gene fusions potentiate Wnt signals. TCF4 mutations are surprisingly common,9 but their roles and effects are unclear, as is the functional significance of rare AXIN2 mutations in MSI-positive cases and of TCF3 or TCF4 gene fusions in microsatellite stable (MSS) cases.

Even advanced CRCs seem to rely on constitutive Wnt pathway activity, leading to much interest in developing therapies that attenuate this pathway, but two prominent barriers prevail. First is the obvious risk of on-target toxicity toward the millions of normal Wnt-dependent colonic cells. Second, APC and CTNNB1 mutations act far downstream of cell surface receptor activity; this restricts potential vulnerabilities to distal points in the signaling pathway, such as stability of TCF–β-catenin complexes41 or their transcriptional targets. Although CD44 and some other target genes, especially MYC, are essential components of the Wnt response in mouse intestines,42 the candidacy and roles of most target genes in human CRC are uncertain. RSPO and RNF43 mutations offer greater promise for treatment because the corresponding CRCs depend in principle on Wnt ligand activity. Indeed, preclinical models show increased sensitivity of RNF43-mutant CRCs to inhibitors of PORCN43 and of tumors with RSPO fusions to specific RSPO antibodies.44

Наследственный неполипозный колоректальный рак и роль ДНК мисматч репарации

HNPCC (Lynch syndrome) is an autosomal dominant disorder that confers 40% to 70% lifetime risk of developing CRC, usually before age 50 years, and accounts for up to 4% of cases. Individuals with HNPCC develop many fewer polyps than patients with FAP, the condition that must be excluded to meet the diagnostic criteria for HNPCC (Table 61.3).45 Cancers tend to arise in the ascending colon, and patients are also prone to develop tumors of the endometrium, ovary, stomach, small bowel, biliary tract, urothelium, and brain (see Table 61.3). The lifetime risk of endometrial cancer, in particular, is 35% to 50% and that of urologic and ovarian tumors is 7% to 8%. The key molecular feature of HNPCC tumors is pronounced variation in the lengths of microsatellite DNA sequences (MSI-hi), which is typically assessed in a panel of five mono- and dinucleotide tracts (BAT26, BAT25, D5S346, D2S123, and D17S250).

Таблица 61.3. Критерии клинической диагностики наследственного неполипозного колоректального рака

A.      Пересмотренные Амстердамские критерии (клинический диагноз)

  1. Three or more family members with histologically verified HNPCC-related cancers, one of whom is a first-degree relative of the other two
  2. Two successive affected generations
  3. One or more of the HNPCC-related cancers (see C) diagnosed before age 50 y
  4. Exclusion of familial adenomatous polyposis
B.      Revised Bethesda guidelines (criteria to prompt MSI testing of tumors)

  1. Diagnosis of CRC before age 50 y
  2. Synchronous or metachronous presence of CRC or other HNPCC-associated cancer
  3. CRC diagnosed before age 60 y with histopathologic features associated with MSI-hi
  4. CRC in at least one first-degree relative with an HNPCC-related tumor, with one of the cancers diagnosed before age 50 y
  5. CRC in two or more first-degree relatives with HNPCC-related tumors, regardless of age
C.      Spectrum of sites for HNPCC-related cancers

Colon and rectum, endometrium, stomach, ovary, pancreas, ureter and renal pelvis, biliary tract, small intestine, brain, sebaceous gland adenomas, and keratoacanthomas

HNPCC — наследственный неполипозный колоректальный рак; MSI-hi — высокая микросателлитная нестабильность; CRC — колоректальный рак.

HNPCC results from germline mutations in any of several genes that enable DNA MMR, the proofreading process that corrects base-pair mismatches and short indels that arise in the normal course of DNA replication. MMR is an efficient process mediated by homologs of bacterial and yeast repair proteins: MutS homologs (MSH) 1 to 6, MutL homologs (MLH) 1 to 3, PMS1, and PMS2. At sites of DNA mismatch, MLH1 and PMS2 are recruited as a MutLα complex; in turn, they recruit MSH2 and MSH6 heterodimers (MutSα) to sites of 1-bp errors and MSH2 and MSH3 (MutSβ) to sites of 2- to 4-bp errors. These complexes excise the strand that carries the mismatch and then resynthesize and ligate the repaired DNA. Germline mutations in MSH2, MLH1, MSH6 and PMS2 together explain about 95% of kindreds,46–48 including a subset with germline loss of the stop codon of EPCAM (previously called TACSTD1), which silences the neighboring MSH2 gene promoter by hypermethylation.49,50

MSI-hi colon cancers are usually exophytic, with medullary histology, lymphocyte infiltrates, and mucinous differentiation; the Revised Bethesda guidelines (see Table 61.3) combine clinical and phenotypic features to facilitate a diagnosis of HNPCC. When these criteria are met, tumor DNA should be tested either for MSI in a simple, polymerase chain reaction (PCR)–based assay or by immunohistochemistry for absence of the most commonly implicated proteins: MLH1, MSH2, and MSH6. Because clinical guidelines might miss up to a quarter of cases, some experts recommend testing all CRCs in patients younger than age 70 years.51,52 Coupled with thorough personal and family histories, positive results should prompt DNA testing for MLH1, MSH2, MSH6, or PMS2 mutations because precise identification of mutant alleles and carriers allows targeted screening and intervention to reduce mortality—colonoscopy every 1 to 2 years starting around age 30 years, family counseling, and aspirin prophylaxis. Two large expert groups have provided consensus guidelines for carriers, including consideration of preventive subtotal colectomy, annual endometrial evaluation for women older than age 30 years, and hysterectomy and oophorectomy after bearing children.52,53

In incipient cancers, random events first disrupt function of the wild-type allele of a mutant MMR gene, resulting in a “mutator phenotype” that increases errors in DNA replication 102- to 103-fold over background rates.6,46 Consequently, adenomas progress into carcinomas over 3 to 5 years instead of two or more decades. The coding regions of the most commonly inactivated genes, ACVR2A and TGFBR2, encode receptors for TGF-β ligands and contain vulnerable mononucleotide tracts.9,54 TGF-β inhibits intestinal epithelial cell proliferation, and >90% of MSI-hi and 15% of MSS, sporadic CRCs show biallelic TGFBR2 inactivation.55 Other genes mutated in familial MSI-hi colon tumors encode the negative Wnt regulator RNF43,56 the pro-apoptotic genes CASP5 and BAX, EGFR, and transcription factor genes, including TCF7L2, but KRAS and especially BRAF mutations are rare in familial cases (Lynch syndrome). CRC pathogenesis appears to require Wnt pathway activation irrespective of MMR status.

MSI-hi is observed in up to 15% of sporadic cases of CRC, often in older patients with early-stage disease in the ascending colon. Such tumors often arise in sessile serrated adenomas, with MLH1 inactivated by biallelic promoter hypermethylation, and they commonly show activating BRAF mutations and CIMP. If BRAF is not mutated, the prognosis for patients with familial or sporadic MSI-hi CRCs is better than for those with sporadic MSS disease. One possible reason is that some somatic mutations limit tumor viability, but observations that CRC prognosis correlates with the extent of lymphocyte infiltration57 also imply a role for adaptive immune responses. Indeed, MSI-hi CRCs are uniquely sensitive to treatment with immune checkpoint inhibitors,58 likely reflecting expression of neoantigens owing to hypermutation, and drugs such as pembrolizumab and nivolumab have a demonstrable role in treating advanced MSI-hi CRC.

Другие наследственные синдромы с повышенным риском колоректального рака

MYH-associated polyposis (MAP), another recessively inherited syndrome of multiple adenomas and CRC, results from germline mutations in MYH, a homolog of the Escherichia coli base excision-repair gene MutY.59 CRCs develop later in life than they do in FAP or HNPCC, polyp numbers vary widely, and extracolonic tumors are uncommon. MYH encodes a DNA glycosylase that mediates oxidative DNA damage; accordingly, tumors are not associated with MSI but with somatic G:C to T:A mutations in genes such as APC. Two alleles, Y165C and G382D, account for most cases, and cancers develop in homozygote or compound heterozygote individuals but not in monoallelic carriers. Clinical findings in PPAP—large adenomas, early-onset CRC, and elevated endometrial cancer risk in women with mutant POLD1—resemble those in HNPCC or MAP. POLE and POLD1 are nonclassical tumor suppressors: The wild-type allele is usually retained, and instead of deletion or truncation, specific missense mutations affect proofreading function.60

Семейный ювенильный полипоз

Patients with familial juvenile polyposis develop premalignant hamartomatous polyps in the stomach, small intestine, or large intestine by adolescence, and a significant minority of cases represent the first in that kindred. Highlighting the role of TGF-β signaling in disease pathogenesis, the genetic basis is germline mutations in genes encoding the BMP receptor BMPR1A, the accessory TGF-β receptor endoglin (ENG), or the signal transducer SMAD461,62; additional genes remain undiscovered. Patients with Peutz-Jeghers syndrome develop benign tumors that contain differentiated but disorganized cells (hamartomas), mainly in the small intestine but also in the colon. This autosomal dominant condition is associated with macular lesions on the skin and buccal mucosa, bladder and bronchial polyps, and a 90% lifetime risk to develop diverse cancers; the incidence of small intestine, stomach, and pancreas cancers is 50 to 500 times higher than the general population, and CRC risk is elevated nearly 100-fold. The implicated tumor suppressor gene, serine–threonine kinase 11 (STK11, also known as LKB1),63 acts at the nexus of diverse cellular pathways and functions, with a principal role in adenosine monophosphate–activated protein kinase (AMPK)–mediated control of nutrient and energy utilization, cellular structure, and apicobasal polarity. STK11 also modulates the Rheb-GDP:Rheb-GTP cycle and downstream activities of the tuberous sclerosis gene TSC2 and the mammalian target of rapamycin (mTOR),64 key regulators of protein synthesis and growth.

The Cowden syndrome encompasses diverse mucosal lesions, cutaneous papules, lipomas, neurofibromas, breast fibroadenomas, and meningiomas. It results from germline mutations in the tumor suppressor PTEN,65 the second most frequently mutated gene in cancers, after TP53. PTEN is a lipid phosphatase that dephosphorylates key phosphoinositide signaling molecules66 to regulate intracellular growth signaling negatively through phosphatidylinositol 3-kinase (PI3K) and its downstream effectors AKT and mTOR. CRC risk in Cowden syndrome is modest and PTEN mutations or deletions occur in 10% of sporadic cases, but the protein is lost in approximately 40% of CRCs, often as a result of promoter hypermethylation.

Понимание от Менделевских синдромов, исследований по всей геномной ассоциации и от микробиома

Following clinical suspicion or diagnosis of a Mendelian risk syndrome, probands and family members should be tested for pertinent germline mutations, receive genetic counseling, and enter programs for cancer prevention and screening. The corresponding molecular defects profoundly inform understanding of sporadic CRC, revealing in particular the seminal role of Wnt signaling and early, rate-limiting effects of APC inactivation or CTNNB1 activation. STK11 and PTEN loss in inherited and sporadic CRCs also shed light on crucial molecular pathways, whereas HNPCC and PPAP help classify the disease and reveal the broader significance of features such as MSI- hi. The cancer spectrum in HNPCC or PPAP and the specific predilection for CRC, however, remain unexplained. Colonic, endometrial, and certain other epithelia may be especially sensitive to mutations that occur in the setting of defective DNA MMR and base-excision repair, loss of the wild-type tumor suppressor allele may occur more readily in these tissues, or they may lack repair safeguards that protect other cell types.

Even in the absence of a recognized predisposition syndrome, individuals with a history of CRC in a first- degree relative are up to four times more likely to develop CRC than those without a family history. Specific environmental factors that compound the risk of developing CRC are complex and insufficiently characterized but include obesity, excessive consumption of red meat, physical inactivity, and vitamin D deficiency. Many of these factors converge on insulin signaling, suggesting a possibly seminal role for insulin-like growth factors in CRC.67 Longstanding inflammatory bowel disease, especially ulcerative colitis, elevates CRC risk up to 10-fold, likely reflecting increased mutation in the setting of repeated mucosal injury and repair. Colitis-associated CRCs often arise within flat adenomatous plaques and areas of nonadenomatous dysplasia. Compared to sporadic cases, APC inactivation is less frequent, TP53 mutations occur earlier in the cancer sequence, and methylation of the p16INK4a tumor suppressor gene is more common.

The quarter or more of sporadic CRC cases with a familial component probably have diverse molecular etiologies, with low risk conferred by some common genetic variants and environmental factors. To date, genome- wide association studies (GWASs) have uncovered statistical associations with at least 20 distinct loci that individually confer small increases in the risk of developing CRC. Frequencies of risk alleles range from <10% to >50% of humans, and each allele elevates CRC risk no more than 7% to 25% above the background in persons with the nonrisk allele.68,69 Even if homozygosity at some loci and additive effects compound this risk, allele frequencies limit the cumulative risk to <50% to 250% over the background and all identified risk variants together explain <5% to 7% of cases with a family history. Although it is not yet possible to predict individuals’ CRC risk or alter screening recommendations based on GWAS genotypes, identification of risk loci is useful for understanding disease determinants. The causal significance of most DNA sequence variants is unclear, but growing evidence indicates that they are quantitative trait loci, representing the cis-regulatory elements for nearby genes, such as MYC, SMAD7, and BMP4. Risk alleles thus affect the expression levels of linked genes, either promoting transformation of normal ISC at a low frequency or, more likely, influencing the oncogenic potential of other events.

Up to one-sixth of CRCs, especially MSI-hi cases, carry DNA and RNA from the invasive anaerobic microorganism Fusobacterium nucleatum,70,71 and a worse prognosis than cases without that DNA. F. nucleatum is an oral microbe that uses its Fap2 lectin to bind Gal-GalNAc moieties on colonic adenoma or cancer cell surfaces and a unique adhesin, FadA to increase Wnt pathway activity. Although F. nucleatum triggers local inflammation and modulates T-cell infiltrates, its causal role remains uncertain. Notably, its DNA persists in distant metastases and experimental xenografts, raising the provocative prospect of antibiotics as adjuncts to cytotoxic or targeted agents in CRC treatment.72

Мутации онкогенов и опухоль-супрессорных генов в прогрессировании колоректального рака

Building on a foundation of constitutive Wnt activity, somatic mutations in oncogenes and tumor suppressor genes confer malignant properties. Recurring mutations provide crucial clues to decipher oncogenic signaling circuits and develop rational therapies (see Table 61.1). The high collective frequency of recurrent KRAS, BRAF, and PIK3CA mutations places EGFR and its downstream effectors extracellular signal-regulated kinases (ERKs, also known as mitogen-activated protein kinases [MAPKs]) at the center of research and therapeutic efforts.

Фигура 61.3. Сигнальные пути, онкогенные мутации и терапевтические возможности при колоректальном раке (CRC). Полезно рассмотреть общие генетические альтерации в CRC в свете общей канонической схемы сигналинга через рецепторные тирозинкиназы, среди которых рецептор эндотелиального фактора роста является ярким примером. KRAS, онкоген, мутируемый в 40% CRC, сигнализирует о рецепторной активации через RAF протеины (включая BRAF, который мутирует в 5-8% CRC) и фосфатидилинозитол-3-киназу (PI3K), каталитическую субъединицу PIK3CA, которая мутирует в 15-20% CRC. Эти трансдукторы, в свою очередь, активируют внутриклеточные пути митоген- активируемой протеинкиназы (MAPK) и AKT или mTOR пути (mammalian target of rapamycin) соответственно.

Hence, common mutations confer growth factor independence on cells, resulting in dysregulated proliferation, protein synthesis, and metabolism. They also represent promising targets for therapeutic interference with aberrantly activated signaling cascades.

KRAS, BRAF и PIK3CA онкогены

The Ras family of small G-proteins transduces growth factor signals and is aberrantly activated in a variety of cancers. KRAS is mutated in about 40% of CRCs73 and NRAS in another <5% of cases. Mutations in both genes cluster in codon 12 or 13 and less commonly in codon 61. KRAS mutations appear even in lesions of low malignant potential, such as aberrant preadenomatous crypt foci and diminutive polyps, and their frequency increases with lesion size.74 Mutant KRAS alone does not initiate adenomas in mice but, when combined with Apc mutation, accelerates tumor progression75,76; removing it from CRC cells or xenografts impedes cell growth. Among the many growth factor signals that KRAS transduces in diverse tissues, its activity in the colonic epithelium and CRC relates most to EGFR. Mutant KRAS locks EGFR signaling in the “on” state; EGFR antibodies are consequently ineffective in this subset of CRCs,12 and targeted therapies will need to disrupt signals further downstream.

As KRAS is a plasma membrane-associated signal transducer for receptor tyrosine kinases, mutant forms potentially deregulate several effector pathways for cell survival, proliferation, and invasion (Fig. 61.3). KRAS signaling recruits RAF kinases to the plasma membrane and triggers MAPK kinase 1 (MEK1) and MAPK kinase 2 (MEK2) to activate ERK1 and ERK2. KRAS mutation thus induces constitutive phosphorylation of ERKs,77 which in turn phosphorylate proteins that control the G1 to S cell cycle transition, among other substrates.78 Although other, non–KRAS-mediated growth factor pathways also activate the MEK/ERK cascade, signaling in CRC is most often deregulated through activating mutations in KRAS or BRAF. The latter gene is mutated in about 10% of CRCs, especially those associated with MSI and CIMP.79,80 V600E, the most common BRAF mutation in CRC, melanoma, and other cancers, affects a residue within the activation loop of the kinase domain and constitutively activates kinase function, probably by several hundredfold and acting as a phosphomimetic.81 Activated BRAF also leads to ERK phosphorylation, releasing intrinsic restraints on cell growth. Indeed, KRAS and BRAF mutations are mutually exclusive in CRC,9,79 suggesting that they represent alternative routes to the same signaling end.

The prognosis in advanced BRAF-mutant CRC, however, is worse than in KRAS-mutant disease, which implies distinctive molecular or cellular features. Moreover, whereas Kras activation in the mouse intestine has modest consequence on its own, Braf V600E expression rapidly induces persistent generalized hyperplasia with a high penetrance of crypt dysplasia, serrated morphology, and invasive MSI-hi cancers with Wnt and ERK pathway deregulation82; in humans, BRAF mutation is a hallmark of nonfamilial MSI-hi CRC and occurs early in sessile serrated adenomas. Notably, whereas BRAFV600E mutant melanomas respond to selective, ATP-competitive BRAF inhibitors such as vemurafenib and take months to manifest secondary resistance, CRCs carrying the same mutation are intrinsically resistant. This is because BRAF inhibition rapidly induces feedback EGFR signaling through KRAS and CRAF, restoring stimuli for CRC replication; melanoma cells avoid such feedback activation because they hardly express EGFR.83,84 In principle, BRAF inhibition should render cells newly sensitive to direct antagonism of EGFR, but in practice combined antagonism of BRAF and EGFR signaling shows limits that are unexplained.

In addition to the ERKs, KRAS also transmits growth signals through PI3K,77,85 which phosphorylates the intracellular lipid PI-4,5-bisphosphate at the 3 position, triggering a cascade that promotes cell survival and growth. Up to 20% of CRCs carry activating mutations in PIK3CA, the gene encoding the catalytic p110 subunit of PI3K; many fewer CRCs carry related PIK3R1 mutations. PIK3CA mutations cluster in exons 9 and 20 and generally arise late in the adenoma–carcinoma sequence, possibly coincident with invasion. Cellular PI3K activity is antagonized by the product of the PTEN gene, which is inactivated—usually by deletion—in another 10% of cases and, as noted previously, is the causal factor in Cowden syndrome. Although both PI3K and BRAF act downstream of KRAS, only BRAF and KRAS mutations are mutually exclusive, and up to one-fifth of KRAS- mutant CRCs also carry PIK3CA mutations, implying that these oncogenes are not totally redundant. One reason could be that mutant KRAS activates PI3K signaling inefficiently. More likely, oncogenic signaling pathways are less strictly linear than is convenient to depict. Indeed, overtly parallel streams of KRAS signaling through RAF- MEK and PI3K (see Fig. 61.3) interact extensively with one another, and both streams feed into mTOR, which coordinates cell growth with nutrient responses.86 Lastly, insulin-like growth factor 2 (IGF2) is overexpressed in approximately 15% of CRCs as a result of focal gene amplifications, loss of imprinting, or other mechanisms.9 IGF2 overexpression is mutually exclusive with genomic events that enhance PI3K signaling, such as PIK3CA mutations and PTEN deletions, suggesting that in normal colon cells, PI3K transmits growth signals from both EGFR and IGF2.

MYC, CDK8 и контроль роста и метаболизма клеток

Although the MYC and CDK8 oncogenes are rarely mutated in CRC, considerable gene amplification is present in approximately 10% of cases and moderately increased copy number and expression are seen in up to 25% of cases9,87; aberrant MYC regulation likely explains the significant GWAS risk allele rs6983267. MYC is not only a prominent target of Wnt signaling,88 but seems to account for the bulk of tumor effect in Apc-mutant mouse intestines.42 CDK8, a cyclin-dependent kinase component of the Mediator complex, couples transcription factors to the basal transcriptional machinery and cooperates with MYC to regulate thousands of genes, including those necessary for cellular metabolism, proliferation, and self-renewal. CDK8 activity in CRC is particularly associated with β-catenin.87 Indeed, whereas APC, CTNNB1, and probably RSPO mutations kick-start adenomas, additional genetic events in CRC potentiate Wnt activity. Disrupting this seminally important pathway and/or its downstream effector MYC may therefore be imperative in CRC therapy but poses formidable challenges, in part because that requires interfering with protein–protein interactions downstream of conventional “druggable” nodes.41

TP53 и другие опухолевые супрессоры

Allelic loss of chromosome 17p is observed in approximately 75% of CRCs but <10% of polyps,74 indicating that it is a late event that may favor tumor progression. In most tumors with this LOH, the remaining TP53 allele is inactivated, most often at codon 175, 245, 248, 273, or 282. Cells with intact TP53 function undergo cell cycle arrest and apoptosis when faced with stress from DNA damage, hypoxia, reduced nutrient access, or aneuploidy. TP53 loss allows cells to overcome these barriers to tumor survival and progression but does not confer specific disease features in CRC. FBXW7, another gene frequently inactivated in CRC, encodes a receptor subunit of Skp, Cullin, F-box-containing (SCF)–E3 ubiquitin ligase complexes, which degrade multiple regulators of cell growth, such as MYC and JUN transcription factors. Monoallelic missense mutations tend to cluster in arginine residues within a β-propeller domain that recognizes specific substrates, including NOTCH, JUN, DEK, and TG- interacting factor 1 (TGIF1) in intestinal cells.

LOH of chromosome 18q is rare in small to midsize adenomas but observed in >60% of CRCs and nearly all liver metastases from MSS tumors. The minimal common region of LOH contains two candidate tumor suppressor genes89: SMAD4 (DPC4) in about one-third of cases and Deleted in Colorectal Cancer (DCC, a receptor for Netrin axonal guidance factors) in the rest. SMAD4/ DPC4 and SMAD2 are positive and negative regulators of TGF-β signaling, respectively, and closely linked on 18q. Somatic SMAD4 mutations are present in 10% to 15% of CRCs with LOH, and germline mutations are noted in some familial juvenile polyposis kindreds; SMAD2 and DCC are rarely mutated in CRC, but DCC messenger RNA (mRNA) and protein are lost in >50% of cases. Together, the findings suggest a complex, multifactorial basis for selection of 18q LOH in CRC.

Прогностическое и предиктивное значение генотипов и молекулярных свойств опухолей

Clinical features and outcomes are similar whether mutations in APC, CTNNB1, or some other gene underlie constitutive Wnt pathway activity. KRAS mutations—and possibly PIK3CA or BRAF mutations or loss of PTEN expression in some contexts—predict lack of response to EGFR antibodies12,90 and thus direct treatment decisions. Mutations in KRAS and PIK3CA seem not to impact survival in stage III or IV disease treated with chemotherapy,13,91 although they will likely predict responses to agents that target MEK or PI3K signaling. Patients with metastatic BRAF-mutant CRC have especially low survival and respond poorly to current chemotherapy regimens, including adjuvant fluoropyrimidines in stage III disease.13,92 Because common CRC mutations have limited prognostic value, attention for this purpose has turned to mRNA expression profiles. Three subgroups defined in one study reflect the distinction among CIN, MSI-hi, and SSA/CIMP tumors,93 whereas others defined up to six subgroups related to “stemness” and “epithelial-mesenchymal transition” phenotypes and variable treatment responses. Shared efforts culminated in delineation of four consensus molecular subtypes (Table 61.4),94 which seem stable in experimental models and can be distinguished in clinical specimens by an immunohistochemistry panel.95 Although this classification may yet find value in patient stratification and treatment decisions, a notable caveat is that its predictive power largely reflects gene expression in stromal rather than tumor cells, with poor-prognosis subtypes revealing a TGF-β–induced stromal cell program.33 Recent studies of organoids cultured from primary CRCs suggest a possible role for personalized in vitro testing to determine which drug combinations may be effective against individual tumors.

Таблица 61.4. Черты, ассоциированные с консенсусными молекулярными подтипами (основанные на экспрессии генов) колоректального рака

CMS1 (иммунный MSI, 14%): гипермутированный (MSI-hi и CIMP), типичны BRAF мутации, инфильтрация активированными лимфоцитами, плохая выживаемость после рецидива

CMS2 (канонический, 37%): высокая частота SCNA, высокая активация Wnt и MYC

CMS3 (метаболический, 13%): низкая частота SCNA или CIMP, типичны KRAS мутации, дерегуляция метаболизма

CMS4 (мезенхимальный, 23%): высокая частота SCNA, инфильтрация стромы, активация TGF-β, ангиогенез, плохая безрецидивная и общая выживаемость

Неклассифицируемый или смешанный (13%)

CMS, consensus molecular subtype; MSI, microsatellite instability; MSI-hi, high microsatellite instability; CIMP, CpG island methylator phenotype; SCNA, somatic copy number alterations; TGF-β, transforming growth factor β.


  1. Jones S, Chen WD, Parmigiani G, et al. Comparative lesion sequencing provides insights into tumor evolution. Proc Natl Acad Sci U S A 2008;105(11):4283–4288.
  2. East JE, Saunders BP, Jass JR. Sporadic and syndromic hyperplastic polyps and serrated adenomas of the colon: classification, molecular genetics, natural history, and clinical management. Gastroenterol Clin North Am 2008;37(1):25–46.
  3. Sjöblom T, Jones S, Wood LD, et al. The consensus coding sequences of human breast and colorectal cancers. Science 2006;314(5797):268–274.
  4. Leary RJ, Lin JC, Cummins J, et al. Integrated analysis of homozygous deletions, focal amplifications, and sequence alterations in breast and colorectal cancers. Proc Natl Acad Sci U S A 2008;105(42):16224–16229.
  5. Bakhoum SF, Ngo B, Laughney AM, et al. Chromosomal instability drives metastasis through a cytosolic DNA response. Nature 2018;553(7689):467–472.
  6. Ionov Y, Peinado MA, Malkhosyan S, et al. Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis. Nature 1993;363(6429):558–561.
  7. Spring KJ, Zhao ZZ, Karamatic R, et al. High prevalence of sessile serrated adenomas with BRAF mutations: a prospective study of patients undergoing colonoscopy. Gastroenterology 2006;131(5):1400–1407.
  8. Sinicrope FA, Rego RL, Halling KC, et al. Prognostic impact of microsatellite instability and DNA ploidy in human colon carcinoma patients. Gastroenterology 2006;131(3):729–737.
  9. The Cancer Genome Atlas Network. Comprehensive molecular characterization of human colon and rectal cancer. Nature 2012;487(7407):330–337.
  10. Seshagiri S, Stawiski EW, Durinck S, et al. Recurrent R-spondin fusions in colon cancer. Nature 2012;488(7413):660–664.
  11. Guda K, Veigl ML, Varadan V, et al. Novel recurrently mutated genes in African American colon cancers. Proc Natl Acad Sci U S A 2015;112(4):1149–1154.
  12. Van Cutsem E, Köhne CH, Hitre E, et al. Cetuximab and chemotherapy as initial treatment for metastatic colorectal cancer. N Engl J Med 2009;360(14):1408–1417.
  13. Souglakos J, Philips J, Wang R, et al. Prognostic and predictive value of common mutations for treatment response and survival in patients with metastatic colorectal cancer. Br J Cancer 2009;101(3):465–472.
  14. Le DT, Durham JN, Smith KN, et al. Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade. Science 2017;357(6349):409–413.
  15. Goelz SE, Vogelstein B, Hamilton SR, et al. Hypomethylation of DNA from benign and malignant human colon neoplasms. Science 1985;228(4696):187–190.
  16. Feinberg AP, Gehrke CW, Kuo KC, et al. Reduced genomic 5-methylcytosine content in human colonic neoplasia. Cancer Res 1988;48(5):1159–1161.
  17. Toyota M, Ahuja N, Ohe-Toyota M, et al. CpG island methylator phenotype in colorectal cancer. Proc Natl Acad Sci U S A 1999;96(15):8681–8686.
  18. Suzuki H, Watkins DN, Jair KW, et al. Epigenetic inactivation of SFRP genes allows constitutive WNT signaling in colorectal cancer. Nat Genet 2004;36(4):417–422.
  19. Xu Y, Hu B, Choi AJ, et al. Unique DNA methylome profiles in CpG island methylator phenotype colon cancers. Genome Res 2012;22(2):283–291.
  20. Mathur R, Alver BH, San Roman AK, et al. ARID1A loss impairs enhancer-mediated gene regulation and drives colon cancer in mice. Nat Genet 2017;49(2):296–302.
  21. Moser AR, Pitot HC, Dove WF. A dominant mutation that predisposes to multiple intestinal neoplasia in the mouse. Science 1990;247(4940):322–324.
  22. Barker N, Ridgway RA, van Es JH, et al. Crypt stem cells as the cells-of-origin of intestinal cancer. Nature 2009;457(7229):608–611.
  23. Shimokawa M, Ohta Y, Nishikori S, et al. Visualization and targeting of LGR5+ human colon cancer stem cells. Nature 2017;545(7653):187–192.
  24. de Sousa e Melo F, Kurtova AV, Harnoss JM, et al. A distinct role for Lgr5+ stem cells in primary and metastatic colon cancer. Nature 2017;543(7647):676–680.
  25. Vermeulen L, Morrissey E, van der Heijden M, et al. Defining stem cell dynamics in models of intestinal tumor initiation. Science 2013;342(6161):995–998.
  26. Carmon KS, Gong X, Lin Q, et al. R-spondins function as ligands of the orphan receptors LGR4 and LGR5 to regulate Wnt/beta-catenin signaling. Proc Natl Acad Sci U S A 2011;108(28):11452–11457.
  27. de Lau W, Barker N, Low TY, et al. Lgr5 homologues associate with Wnt receptors and mediate R-spondin signalling. Nature 2011;476(7360):293–297.
  28. Koo BK, Spit M, Jordens I, et al. Tumour suppressor RNF43 is a stem-cell E3 ligase that induces endocytosis of Wnt receptors. Nature 2012;488(7413):665–669.
  29. de Lau W, Peng WC, Gros P, et al. The R-spondin/Lgr5/Rnf43 module: regulator of Wnt signal strength. Genes Dev 2014;28(4):305–316.
  30. Haramis AP, Begthel H, van den Born M, et al. De novo crypt formation and juvenile polyposis on BMP inhibition in mouse intestine. Science 2004;303(5664):1684–1686.
  31. He XC, Zhang J, Tong WG, et al. BMP signaling inhibits intestinal stem cell self-renewal through suppression of Wnt-β-catenin signaling. Nat Genet 2004;36(10):1117–1121.
  32. Batts LE, Polk DB, Dubois RN, et al. Bmp signaling is required for intestinal growth and morphogenesis. Dev Dyn 2006;235(6):1563–1570.
  33. Calon A, Lonardo E, Berenguer-Llergo A, et al. Stromal gene expression defines poor-prognosis subtypes in colorectal cancer. Nat Genet 2015;47(4):320–329.
  34. Drost J, van Jaarsveld RH, Ponsioen B, et al. Sequential cancer mutations in cultured human intestinal stem cells. Nature 2015;521(7550):43–47.
  35. Fujii M, Shimokawa M, Date S, et al. A colorectal tumor organoid library demonstrates progressive loss of niche factor requirements during tumorigenesis. Cell Stem Cell 2016;18(6):827–838.
  36. Roper J, Tammela T, Cetinbas NM, et al. In vivo genome editing and organoid transplantation models of colorectal cancer and metastasis. Nat Biotechnol 2017;35(6):569–576.
  37. Powell SM, Zilz N, Beazer-Barclay Y, et al. APC mutations occur early during colorectal tumorigenesis. Nature 1992;359(6392):235–237.
  38. Sansom OJ, Reed KR, Hayes AJ, et al. Loss of Apc in vivo immediately perturbs Wnt signaling, differentiation, and migration. Genes Dev 2004;18(12):1385–1390.
  39. Morin PJ, Sparks AB, Korinek V, et al. Activation of beta-catenin-Tcf signaling in colon cancer by mutations in beta-catenin or APC. Science 1997;275(5307):1787–1790.
  40. Sparks AB, Morin PJ, Vogelstein B, et al. Mutational analysis of the APC/beta-catenin/Tcf pathway in colorectal cancer. Cancer Res 1998;58(6):1130–1134.
  41. Lepourcelet M, Chen YN, France DS, et al. Small-molecule antagonists of the oncogenic Tcf/beta-catenin protein complex. Cancer Cell 2004;5(1):91–102.
  42. Sansom OJ, Meniel VS, Muncan V, et al. Myc deletion rescues Apc deficiency in the small intestine. Nature 2007;446(7136):676–679.
  43. Koo BK, van Es JH, van den Born M, et al. Porcupine inhibitor suppresses paracrine Wnt-driven growth of Rnf43;Znrf3-mutant neoplasia. Proc Natl Acad Sci U S A 2015;112(24):7548–7550.
  44. Storm EE, Durinck S, de Sousa e Melo F, et al. Targeting PTPRK-RSPO3 colon tumours promotes differentiation and loss of stem-cell function. Nature 2016;529(7584):97–100.
  45. Vasen HF, Watson P, Mecklin JP, et al. New clinical criteria for hereditary nonpolyposis colorectal cancer (HNPCC, Lynch syndrome) proposed by the International Collaborative group on HNPCC. Gastroenterology 1999;116(6):1453–1456.
  46. Fishel R, Kolodner RD. Identification of mismatch repair genes and their role in the development of cancer. Curr Opin Genet Dev 1995;5(3):382–395.
  47. Vasen HF, Boland CR. Progress in genetic testing, classification, and identification of Lynch syndrome. JAMA 2005;293(16):2028–2030.
  48. Liu T, Yan H, Kuismanen S, et al. The role of hPMS1 and hPMS2 in predisposing to colorectal cancer. Cancer Res 2001;61(21):7798–7802.
  49. Ligtenberg MJ, Kuiper RP, Chan TL, et al. Heritable somatic methylation and inactivation of MSH2 in families with Lynch syndrome due to deletion of the 3′ exons of TACSTD1. Nat Genet 2009;41(1):112–117.
  50. Kovacs ME, Papp J, Szentirmay Z, et al. Deletions removing the last exon of TACSTD1 constitute a distinct class of mutations predisposing to Lynch syndrome. Hum Mutat 2009;30(2):197–203.
  51. Hampel H, Frankel WL, Martin E, et al. Screening for the Lynch syndrome (hereditary nonpolyposis colorectal cancer). N Engl J Med 2005;352(18):1851–1860.
  52. Vasen HF, Blanco I, Aktan-Collan K, et al. Revised guidelines for the clinical management of Lynch syndrome (HNPCC): recommendations by a group of European experts. Gut 2013;62(6):812–823.
  53. Giardiello FM, Allen JI, Axilbund JE, et al. Guidelines on genetic evaluation and management of Lynch syndrome: a consensus statement by the US Multi-Society Task Force on Colorectal Cancer. Gastroenterology 2014;147(2):502–526.
  54. Markowitz S, Wang J, Myeroff L, et al. Inactivation of the type II TGF-beta receptor in colon cancer cells with microsatellite instability. Science 1995;268(5215):1336–1338.
  55. Grady WM, Myeroff LL, Swinler SE, et al. Mutational inactivation of transforming growth factor beta receptor type II in microsatellite stable colon cancers. Cancer Res 1999;59(2):320–324.
  56. Giannakis M, Hodis E, Jasmine Mu X, et al. RNF43 is frequently mutated in colorectal and endometrial cancers. Nat Genet 2014;46(12):1264–1266.
  57. Galon J, Costes A, Sanchez-Cabo F, et al. Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Science 2006;313(5795):1960–1964.
  58. Le DT, Uram JN, Wang H, et al. PD-1 blockade in tumors with mismatch- repair xeficiency. N Engl J Med 2015;372(26):2509–2520.
  59. Sieber OM, Lipton L, Crabtree M, et al. Multiple colorectal adenomas, classic adenomatous polyposis, and germ- line mutations in MYH. N Engl J Med 2003;348(9):791–799.
  60. Palles C, Cazier JB, Howarth KM, et al. Germline mutations affecting the proofreading domains of POLE and POLD1 predispose to colorectal adenomas and carcinomas. Nat Genet 2013;45(2):136–144.
  61. Howe JR, Sayed MG, Ahmed AF, et al. The prevalence of MADH4 and BMPR1A mutations in juvenile polyposis and absence of BMPR2, BMPR1B, and ACVR1 mutations. J Med Genet 2004;41(7):484–491.
  62. Sweet K, Willis J, Zhou XP, et al. Molecular classification of patients with unexplained hamartomatous and hyperplastic polyposis. JAMA 2005;294(19):2465–2473.
  63. Hemminki A, Markie D, Tomlinson I, et al. A serine/threonine kinase gene defective in Peutz-Jeghers syndrome. Nature 1998;391(6663):184–187.
  64. Shaw RJ, Bardeesy N, Manning BD, et al. The LKB1 tumor suppressor negatively regulates mTOR signaling. Cancer Cell 2004;6(1):91–99.
  65. Liaw D, Marsh DJ, Li J, et al. Germline mutations of the PTEN gene in Cowden disease, an inherited breast and thyroid cancer syndrome. Nat Genet 1997;16(1):64–67.
  66. Maehama T, Dixon JE. The tumor suppressor, PTEN/MMAC1, dephosphorylates the lipid second messenger, phosphatidylinositol 3,4,5-trisphosphate. J Biol Chem 1998;273(22):13375–13378.
  67. Slattery ML, Fitzpatrick FA. Convergence of hormones, inflammation, and energy-related factors: a novel pathway of cancer etiology. Cancer Prev Res (Phila) 2009;2(11):922–930.
  68. Tomlinson I, Webb E, Carvajal-Carmona L, et al. A genome-wide association scan of tag SNPs identifies a susceptibility variant for colorectal cancer at 8q24.21. Nat Genet 2007;39(8):984–988.
  69. Tenesa A, Farrington SM, Prendergast JG, et al. Genome-wide association scan identifies a colorectal cancer susceptibility locus on 11q23 and replicates risk loci at 8q24 and 18q21. Nat Genet 2008;40(5):631–637.
  70. Kostic AD, Gevers D, Pedamallu CS, et al. Genomic analysis identifies association of Fusobacterium with colorectal carcinoma. Genome Res 2012;22(2):292–298.
  71. Castellarin M, Warren RL, Freeman JD, et al. Fusobacterium nucleatum infection is prevalent in human colorectal carcinoma. Genome Res 2012;22(2):299–306.
  72. Bullman S, Pedamallu CS, Sicinska E, et al. Analysis of Fusobacterium persistence and antibiotic response in colorectal cancer. Science 2017;358(6369):1443–1448.
  73. Bos JL, Fearon ER, Hamilton SR, et al. Prevalence of ras gene mutations in human colorectal cancers. Nature 1987;327(6120):293–297.
  74. Vogelstein B, Fearon ER, Hamilton SR, et al. Genetic alterations during colorectal-tumor development. N Engl J Med 1988;319(9):525–532.
  75. Sansom OJ, Meniel V, Wilkins JA, et al. Loss of Apc allows phenotypic manifestation of the transforming properties of an endogenous K-ras oncogene in vivo. Proc Natl Acad Sci U S A 2006;103(38):14122–14127.
  76. Feng Y, Bommer GT, Zhao J, et al. Mutant KRAS promotes hyperplasia and alters differentiation in the colon epithelium but does not expand the presumptive stem cell pool. Gastroenterology 2011;141(3):1003–1013.e10.
  77. Ebi H, Corcoran RB, Singh A, et al. Receptor tyrosine kinases exert dominant control over PI3K signaling in human KRAS mutant colorectal cancers. J Clin Invest 2011;121(11):4311–4321.
  78. Downward J. Targeting RAS signalling pathways in cancer therapy. Nat Rev Cancer 2003;3(1):11–22.
  79. Rajagopalan H, Bardelli A, Lengauer C, et al. Tumorigenesis: RAF/RAS oncogenes and mismatch-repair status. Nature 2002;418(6901):934.
  80. Tanaka H, Deng G, Matsuzaki K, et al. BRAF mutation, CpG island methylator phenotype and microsatellite instability occur more frequently and concordantly in mucinous than non-mucinous colorectal cancer. Int J Cancer 2006;118(11):2765–2771.
  81. Davies H, Bignell GR, Cox C, et al. Mutations of the BRAF gene in human cancer. Nature 2002;417(6892):949– 954.
  82. Rad R, Cadiñanos J, Rad L, et al. A genetic progression model of Braf(V600E)- induced intestinal tumorigenesis reveals targets for therapeutic intervention. Cancer Cell 2013;24(1):15–29.
  83. Corcoran RB, Ebi H, Turke AB, et al. EGFR-mediated re-activation of MAPK signaling contributes to insensitivity of BRAF mutant colorectal cancers to RAF inhibition with vemurafenib. Cancer Discov 2012;2(3):227–235.
  84. Prahallad A, Sun C, Huang S, et al. Unresponsiveness of colon cancer to BRAF(V600E) inhibition through feedback activation of EGFR. Nature 2012;483(7387):100–103.
  85. Gupta S, Ramjaun AR, Haiko P, et al. Binding of ras to phosphoinositide 3-kinase p110alpha is required for ras-driven tumorigenesis in mice. Cell 2007;129(5):957–968.
  86. Shaw RJ, Cantley LC. Ras, PI(3)K and mTOR signalling controls tumour cell growth. Nature 2006;441(7092):424–430.
  87. Firestein R, Bass AJ, Kim SY, et al. CDK8 is a colorectal cancer oncogene that regulates beta-catenin activity. Nature 2008;455(7212):547–551.
  88. van de Wetering M, Sancho E, Verweij C, et al. The beta-catenin/TCF-4 complex imposes a crypt progenitor phenotype on colorectal cancer cells. Cell 2002;111(2):241–250.
  89. Thiagalingam S, Lengauer C, Leach FS, et al. Evaluation of candidate tumour suppressor genes on chromosome 18 in colorectal cancers. Nat Genet 1996;13(3):343–346.
  90. Sartore-Bianchi A, Di Nicolantonio F, Nichelatti M, et al. Multi-determinants analysis of molecular alterations for predicting clinical benefit to EGFR-targeted monoclonal antibodies in colorectal cancer. PLoS One 2009;4(10):e7287.
  91. Ogino S, Meyerhardt JA, Irahara N, et al. KRAS mutation in stage III colon cancer and clinical outcome following intergroup trial CALGB 89803. Clin Cancer Res 2009;15(23):7322–7329.
  92. Ogino S, Shima K, Meyerhardt JA, et al. Predictive and prognostic roles of BRAF mutation in stage III colon cancer: results from intergroup trial CALGB 89803. Clin Cancer Res 2012;18(3):890–900.
  93. De Sousa e Melo F, Wang X, Jansen M, et al. Poor-prognosis colon cancer is defined by a molecularly distinct subtype and develops from serrated precursor lesions. Nat Med 2013;19(5):614–618.
  94. Guinney J, Dienstmann R, Wang X, et al. The consensus molecular subtypes of colorectal cancer. Nat Med 2015;21(11):1350–1356.
  95. Trinh A, Trumpi K, De Sousa e Melo F, et al. Practical and robust identification of molecular subtypes in colorectal cancer by immunohistochemistry. Clin Cancer Res 2017;23(2):387–398.
  96. Vlachogiannis G, Hedayat S, Vatsiou A, et al. Patient-derived organoids model treatment response of metastatic gastrointestinal cancers. Science 2018;359(6378):920–926.