Reactive oxygen species derived from normal aerobic metabolism and exogenous sources readily damage DNA forming a wide variety of lesions, including oxidized bases [171–174]. One well-studied and common oxidized base is 7,8-dihydro-8-oxo-deoxyguanosine (8-oxo-G). 8-oxo-G readily pairs with adenine, leading to G:C → T:A transversion mutations following DNA replication [172–174]. MYH, the MUTYH gene product, is a glycosylase that catalyzes the cleavage of adenine mismatched to 8-oxo-G via base excision repair, preventing G:C → T:A transversions [175, 176]. Biallelic MUTYH mutation results in glycosylase activity loss and increased G:C → T:A transversions, including in APC and KRAS, but not Braf, leading adenoma development and CRC [177, 178].

MAP is the most important differential diagnosis for FAP and AFAP, occurring in ~1/10,000 individuals, accounting for ~0.6 % of all CRC [179, 180]. Phenotypically, biallelic MUTYH mutation (homozygous or compound heterozygous) results in a 93-fold increased CRC risk, with clinical symptoms appearing at 40–60 years, and a variable number of adenomatous colonic polyps, 10–100, rarely up to 1,000 [180]. Extraintestinal manifestations are less common than in FAP, with ~31 % of effected individuals having duodenal adenomas and/or gastric fundic polyps and a 4 % lifetime duodenal cancer risk [180–182]. MAP is autosomal recessive and therefore does not show the strong intergenerational history of FAP [180, 181]. The CRC incidence is high in MAP , with a 100 % lifetime risk and 50 % of effected individuals having CRC at diagnosis [181, 183]. Extraintestinal manifestations of MAP include breast, endometrial, and sebaceous carcinomas [182, 184]. Interesting, neither immunohistochemistry nor histology reveal any significant differences between MAP -associated CRC and sporadic CRC [185].

Unlike the 700+ known APC mutations, about 23 pathogenic MUTYH mutations are known, with two (Y165C and G382D), accounting for ~82 % of inactivating mutations [114, 115, 186]. Thus initial MAP testing is often directed at these two mutations. If one of these mutations is identified, sequencing is performed to identify the inactivating mutation on the opposite allele. If two inactivating mutations are found, the CRC risk will be very high. Failure to initially identify the Y165C or G382D combined with a strong suspicion for MUTYH mutation necessitates testing for less common mutations usually by DNA sequencing, which has the highest clinical sensitivity [187].

Numerous studies have demonstrated that in many cases ARMS is useful in MAP testing. For example, Piccioli et al. [188] employed a tetra-primer amplification refractory mutation system PCR (T-ARMS-PCR) to examine six common MUTYH mutations , which would identify at least 85 % of all mutations; Y165C, G382D, 1395_7delGGA, Y90X, 1103delC, and R231H. T-ARMS-PCR amplifies both wild-type and mutant alleles in one PCR reaction. T-ARMS-PCR and ARMS-PCR work well for MAP testing and cover ~85 % of known mutations [188, 189]. It is relatively inexpensive, can identity small deletions, requires no specialized equipment, and importantly is ideal for targeted molecular testing. Thus, once a kindred-specific mutation has been identified, ARMS-PCR and T-ARMS-PCR allow inexpensive, accurate, and rapid molecular analyses. Again ARMS-PCR does not identify ~15 % of known mutations, making DNA sequencing analysis necessary to identify uncommon mutations [187, 189]. The T-ARMS-PCR protocol is described in Fig. 5.3.

For individuals with a significant colonic adenoma burden, MUTYH and APC mutation testing can be performed simultaneously, particularly if the family history suggests an autosomal recessive inheritance pattern. MUTYH testing should also be considered for the 20–30 % and 70–90 % of APC mutation negative FAP or AFAP cases, respectively. For FAP, the number due to MUTYH mutations is low, accounting for only ~5 % of FAP [111–114]. Once an individual is identified as having inactivating biallelic MUTYH mutations, molecular testing of all siblings is recommended as they have a 25 % chance of having the same mutation [112]. Presently, testing the partners of individuals with bilallelic MUTYH mutations is controversial, as the risk of MAP for children of an effected parent is ~1 % [181]. However if the partner of a bilallelic mutation carrier has one MUTYH mutation, an apparently autosomal dominant CRC inheritance is seen, possibly resulting in a misdiagnosis of FAP. Additionally, the CRC risk for MUTYH mutation heterozygotes is low, with a relative risk of 1.5–2.1 to the general population. Presently, the need for increased lower endoscopic surveillance is this population is undetermined [112, 190, 191].

Individuals with biallelic MUTYH mutations should undergo annual or biennial lower endoscopic surveillance beginning as teens. Upper endoscopic surveillance for duodenal adenomas/adenocarcinomas should be performed every 2 years [192]. The high rate of progression to CRC often necessitates a prophylactic colectomy [112, 192].

MLH1 is located at 3p21.3-23 and consists of 19 exons of 57.36 kb, coding for an 84.6-kDa nuclear protein. MLH1 has two interaction domains, one which interacts with MSH2, MSH3, or MSH6 and the other which interacts with PMS2, MLH3, or PMS1. More than 250 different MLH1 germline mutations in NHPCC have been identified, consisting of missense, nonsense, splicing errors, and insertion/deletions, many of which truncated protein expression [198, 199]. Germline MLH1 mutations are seen in ~46 % of HNPCC cases, while 78 % of sporadic CRCs with high MSI (MSI-H) have reduced MLH1 expression secondary to promoter methylation and BRAF mutation. Thus MLH1 inactivation appears to occur by separate molecular pathways in HNPCC vs. sporadic CRC [200].

MSH2 is located at 2p22-p21 and consists of 16 exons of 80.10 kb, coding for a 104.7 kDa nuclear protein. MSH2 has DNA, MSH3/MSH6, and MLH1/PMS2 binding domains. More than 300 different MLH2 germline mutations in HNPCC have been identified, consisting of mutation types similar to those seen in MHL1. MLH2 mutations are seen in ~40 % of HNPCC cases, thus ~90 % of HNPCC CRCs result from MHL1 or MLH2 mutations [198–202].

MSH6 is located at 2p16 and consists of 10 exons of 23.8 kb, coding for a 160 kDa nuclear protein. MSH6 mutations account for 7–10 % of HNPCC [202–204]. 73 % of female MSH6 mutation carriers develop endometrial hyperplasia and carcinoma, compared to 29 % in MSH2 and 31 % in MLH1 carriers, making endometrial tumors more common among female MSH6 mutation carriers than CRC [205]. Most mutations are truncating mutations producing an inactive protein [205]. Interestingly, individuals with MSH6 mutations have a lower CRC incidence and a later age of onset than those with MSH2 and MLH1mutations [206].

PMS2 is located at 7p22.2 and consists of 15 exons of 63.14 kb, coding for a 96 kDa nuclear protein. PMS2 mutations account for less than 2 % of HNPCC kindreds and normally function as heterodimers with MLH1 forming MutLα [198, 202–204]. MSH2, MSH6, MLH1, and PMS2 associate with BRCA1, ATM, RAD50, MRE11, NBS1, and PCNA, forming a complex suggested to regulate DNA replication and maintain genomic integrity [207].

Wu et al. [208] examined 39 unrelated HNPCC families who fulfilled the Amsterdam Criteria for HNPCC. Of 288 individuals examined, 12 had 10 different germline MLH3 variants, including one frameshift, and nine missense mutations. Normal immunostaining for MSH2, MLH1, or MSH6 was detected. The authors concluded that MLH3 mutation probably played a role in a small subset of HNPCC.

The molecular mechanism underlying CRC carcinogenesis in HNPCC appears to be increased genomic instability in ~32 target genes that have mononucleotide repeats of 7 or more within their coding sequences. Many of these genes, such as RAD50, BLM, and PTEN, are tumor suppressors; others such as TGFβ1R2 and BAX inhibit epithelial cell growth or exert proapoptotic effects. MSI causes frameshift mutations in these genes, leading to loss of function [209, 210]. Interestingly, MLH3, MSH3, and MSH6 have A₉, A₈, and C₈ tracts, respectively, thus loss of one MMR gene may lead to progressive inactivation of the entire MMR system, possibly explaining the rapid adenoma → CRC progression seen in HNPCC [211].

Once HNPCC is suspected, specific testing protocols should be followed to obtain a definitive diagnosis. First, MSI or immunohistochemical analyses of CRC and normal tissue should be performed. Should MSI be present, or MMR protein expression lost in CRC, further analyses are performed to diagnose HNPCC [212, 213].

MSI testing involves the analysis of five microsatellite markers standardized by the National Cancer Institute for HNPCC screening: BAT-25, BAT-26, D2S123, D5S346, and D17S250. Typically, both tumor and matching normal patient DNA are compared. MSI testing may be performed without normal patient DNA, but it is recommended that all studies assessing MSI using these markers should use matched normal-tumor DNA pairs [212]. Each microsatellite marker is PCR amplified with primers specific for each marker, and the resulting amplicons are analyzed for length variation via techniques such as gel electrophoresis, automated sizing equipment, or denaturing high-performance liquid chromatography. The primers may be labeled or unlabeled, depending on the requirements of the amplicon detection method [213–215].

MSI is defined as a change of any length due to a deletion/insertion of repeating units within a tumor microsatellite not seen in normal tissue DNA. With five markers, MSI is present and high (MSI-H) when ≥ 2 or more microsatellite markers show MSI. MSI is low (MSI-L) when one marker has MSI, and the tumor is MSI-L or microsatellite stable when one or no markers show MSI. Where more the five markers are analyzed, ≥ 30–40 % with MSI equals MSI-H, < 30–40 % is MSI-L, and no MSI is microsatellite stable [213]. Microsatellite stable or MSI-L tumors are unlikely to result from HNPCC [212]. An example of an MSI analysis is shown in Fig. 5.4.

Immunohistochemical analysis for the loss of MMR genes is as effective as molecular MSI testing for HNPCC [212, 215, 216]. Immunohistochemical analysis of CRC and adjacent normal colorectal epithelium from individuals with HNPCC will often reveal loss of nuclear MSH2, MLH1, MSH6, or PMS2 protein expression. Following CRC analysis, if staining for each MMR protein is identified, the individual is unlikely to have HNPCC. If staining for an MMR protein is lost in CRC, but present in adjacent normal colonic epithelium, HNPCC is likely to be present [212, 215, 216]. The correlation between immunohistochemical MMR protein loss in CRC and MSI is high, with some studies showing MLH1 or MSH2 loss having 100 % specificity and 92.3 % sensitivity for predicting CRC with MSI-H [216]. Although both MSI and immunohistochemistry will work for initial testing, ideally both should be performed [212, 215]. Interestingly, both MSI-H and loss of MMR protein expression has been identified in a significant portion of colonic adenomas from individuals with HNPCC, demonstrating that these events occur early in CRC genesis [215]. Immunohistochemical testing has the advantage of being both cheaper and faster than MSI testing, with costs estimated to be 14–28 % of MSI testing [217]. There is growing interest in the application of reflex testing for MMR and/or MSI for all patients diagnosed with CRC. In particular, the Evaluation of Genomic application in Practice and Prevention (EGAPP) Working Group has provided genetic testing strategies in newly diagnosed individuals with colorectal cancer aimed at reducing morbidity and mortality from Lynch Syndrome in relative.

With no loss of MMR protein immunopositivity, or MSI testing showing MSI-L/MMS, HNPCC is unlikely and no further testing is required [212]. Should MSI-H and/or MMR protein loss be identified, molecular testing should be performed for MMR gene mutations. Most protocols recommend MLH1 and MSH2 molecular testing, as these make up ~90 % of MMR gene HNPCC mutations [198–202, 212, 215, 216]. If immunohistochemistry detected loss of a specific MMR protein, molecular testing should be directed to the encoding gene [212, 215, 216].

Ten to fifteen percent of sporadic CRCs are MSI-H, have MLH1 protein loss, and have the specific BRAF-V600E mutation [82–86, 218, 219]. In these tumors, the MLH1 promoter is hypermethylated, leading to loss of protein expression, while the MLH1 sequence is unmutated [84, 85, 218, 219]. BRAF-V600E occurs only in sporadic CRCs and is not seen in HNPCC-associated CRCs [84, 85, 220, 221]. Therefore, MLH1 molecular testing is unnecessary where a BRAF-V600E mutation has been identified, unless there remains a high clinical suspicion for HNPCC [84, 85, 195, 220, 221]. Thus BRAF-V600E testing can rule out HNPCC, saving time and avoiding costly MMR gene molecular analysis [82, 84, 85, 220].

Molecular testing for BRAF-V600E is described above. Testing typically follows two approaches:

Failure to identify an MMR gene mutation usually rules out HNPCC, while identification of an MMR gene allows an HNPCC diagnosis. In some cases, a gene variation of unknown significance is identified. In other cases, no MMR gene mutation is identified, but the family history and/or individual’s presentation still strongly indicate HNPCC. Genetic testing will not identify all mutations leading to a HNPCC phenotype. Where an MMR gene mutation is identified, directed testing should be performed on the probands at-risk relatives and genetic counseling should be offered [195].

MMR gene analysis has been performed in a variety of ways, including Southern blotting, SSCP, DNA sequencing, MLPA, denaturing gradient gel-electrophoresis, quantitative PCR, and combinations of these methods [222–225]. Many germline MLH1 and MSH2 mutations are missense, nonsense, and frameshift mutations and are readily detected by these techniques. However, other mutations include large deletions and duplications [225–227]. While Southern analysis will detect large genomic rearrangements, it is also cumbersome, expensive, and time-consuming compared to other methods [222]. Charbonnier et al. [225] employed short fluorescent fragments to analyze MLH1 and MSH2 mutations, detecting deletions and duplications. 19 and 16 labeled primer sets were required for MLH1 and MSH2 analysis, respectively, with the reactions multiplexed and analyzed on a Gene scanner model 672 fluorescent fragment Analyzer. Control and patient electropherograms were superimposed and amplifications/deletions were detected by observing changes between superimposed control DNA and control and patient DNA. Most MLH1 and MSH2 mutations are unique with only a few common and recurring mutations identified, thus ARMS would be an extremely inefficient approach to MMR gene mutation analysis [226]. An example of the technique employed in Vasen et al. [228] is shown in Fig. 5.5.

Järvinen et al. [229] compared two at-risk HNPCC cohorts over a 15-year period. Colonoscopic screening reduced the CRC incidence 62 %. Additionally, the CRCs in the screened cohort were localized and caused no deaths, compared to 9 CRC-related deaths in individuals who declined colonoscopy. Interestingly, Vasen et al. [228] found several HNPCC patients with CRC 3.5 years following a negative colonoscopy, demonstrating a need for short screening intervals. Currently, individuals with HNPCC should receive colonoscopy every 1–2 years, beginning at age 20–25 (age 30 years with MSH6 mutation), or 10 years before the first effected family member presented with CRC.

In HNPCC, an increased incidence of CRC is found arising in flat and diminutive adenomas proximal to the splenic flexure. Interestingly, pan-colonic chromoscopic colonoscopy significantly increased flat and pedunculated adenoma detection in HNPCC. Presently however, pan-colonic chromoscopic colonoscopy is not routinely employed in the endoscopic surveillance of individuals with HNPCC, although it may allow a longer surveillance period prior to colectomy [230]. If endoscopic surveillance cannot be performed, a subtotal colectomy should be considered [231]. Prophylactic colectomy, whether subtotal or complete proctocolectomy, is not recommended if annual endoscopic surveillance can be carried out [231]. When unresectable adenomas or CRC are found, subtotal colectomy may be performed due the increased rate of secondary tumors [231].

Women with HNPCC have a 20–60 % lifetime risk for endometrial cancer and should undergo annual endometrial biopsy beginning at age 30–35 years. Transvaginal ultrasound has little value in screening for endometrial cancer, but can be used for annual ovarian screening [231–233]. Prophylactic hysterectomy/oophorectomy can be performed once the childbearing years are completed, as it results in a significant reduction in endometrial and ovarian cancer [234]. Screening for the more common Lynch Syndrome-associated malignancies, such as gastric cancer and urothelial cancer, is also recommended with frequency dictated by individual and family history (NCCN CRC screening guidelines). Screening for rarer malignancies is not defined by guidelines but is determined by individual and family history.

Most studies on NSAID use and chemoprevention in hereditary CRC have focused on FAP . There are no data on the effects of NSAIDS on chemoprevention in HNPCC, and they are currently not recommended. Additionally, although oral contraceptives are known to lower the incidence of endometrial and ovarian cancer in the general population, no data have demonstrated a chemopreventative effect in HNPCC. Last, individuals with HNPCC show an increased CRC with smoking and should be encouraged not to smoke [231].

JPS was first reported by Diamond et al. [235] who described a 30-month-old girl with a prolapsed polyp, accompanied by constipation, hematochezia, and a pedunculated sessile polyp on proctoscopy. In 1975, Stemper et al. [236] described a family with 10 affected members; juvenile polyps were identified in the stomach, small and large intestines, and rectum. Several developed gastric and CRC. Since that time, many studies have found increased gastrointestinal malignancies in kindreds affected by JPS , including gastric, colon, pancreatic, and small intestinal cancers [236–240]. The lifetime risk for malignancy in JPS is estimated to be 9–50 % [241, 242]. Histologically juvenile polyps are characterized by a smooth appearance, with cystically dilated glands embedded in a hyperplasic stroma which has an inflammatory infiltrate.

JPS is presently defined by any of three manifestations: (1) more than 5 juvenile polyps in the colorectum, (2) multiple juvenile polyps throughout the GI tract, and (3) any number of juvenile polyps and a family history of JPS [243]. JPS has been divided into three distinct categories: (1) JPS where the disease is limited to the colon, (2) JPS of infancy characterized by early onset, failure to thrive, hypoalbuminemia, and a poor prognosis, and (3) generalized JPS , where juvenile polyps are seen in the upper and lower gastrointestinal tract [244]. JPS is rare, occurring in ~1 in 10⁵ individuals and shows an autosomal dominant inheritance pattern, with approximately 75 % of individuals having an affected parent and 25 % of cases arising de novo [245, 246]. The diagnosis of JPS is largely clinical, with molecular testing employed for prenatal diagnosis, kindred testing, and clinical diagnosis confirmation.

Two genes are mutated in PJS: BMPR1A and SMAD4 , both accounting for ~20 % of JPS cases [247].

SMAD4 or mothers against decapentaplegic homolog 4, is located at 18q21.1 and consists of 13 exons, which encode a 60.4-kDa protein of 552 residues. SMAD4 encodes an intracellular mediator of the TGF-β family and activin type 1 receptor signaling. Ligation of the TGF-β family receptor leads to SMAD1,−5 and−8 phosphorylation and activation, forming a complex with Smad4, activating pathways that mediate growth inhibitory signals [247–249]. Over 60 SMAD4 mutations have been identified in JPS . Interestingly, one mutation identified in multiple kindreds deletes 4 nucleotides (1244_1247delACAG), creating a truncated protein of 433 residues lacking regions critical for normal protein activity [249, 250].

BMPR1A or bone morphogenetic protein receptor, type IA, is located at 10q22.3, and consists of 13 exons that encode a 60.2 kDA protein of 532 residues. BMPR1A is a serine-threonine kinase, which upon binding TGF-β, phosphorylates and activates SMAD1,−5, and−8. These SMADs bind SMAD4 , creating a complex that inhibits cell growth and division [247–249]. Thus loss of either BMPR1A or SMAD4 activity results in cellular release from the TGF-β growth-inhibitory pathway, increasing malignant transformation risk [247, 248]. More than 60 BMPR1A have been identified, most of which produce a truncated, nonfunctional protein [247–250]. Together, BMPR1A or SMAD4 mutations account for ~45 % of all JPS cases [250]. Patients lacking BMPR1A or SMAD4 mutations are likely to have undiscovered gene mutations. Therefore, failure to identity BMPR1A or SMAD4 mutations does not exclude a JPS diagnosis. The identities of other genes mutated in JPS are currently unknown, although some data implicate endoglin (Eng) gene mutation [251].

Due to the large number of BMPR1A or SMAD4 mutations, DNA sequencing is most often employed in mutation analysis of these genes, although techniques such as PCR based denaturing gradient gel electrophoresis have been employed [249–251]. As for other genes analyzed, a set of specific forward and reverse primers first amplify the genomic DNA, which is purified, analyzed in a commercial sequencer, and compared to the wild-type sequence. Where a mutation is identified, it is often confirmed by sequencing in the opposite direction [249]. Such approaches have been able to identity 18 % and 21 % of BMPR1A and SMAD4 JPS mutations, respectively [249, 250]. As with most direct sequencing techniques, a number of deletions/insertions are missed; hence, techniques such as MLPA have been used to identify 6 % and 7 % more BMPR1A and SMAD4 JPS mutations, respectively [250]. Once identified, directed mutational testing can be performed on individuals related to the proband, who may be at risk for JPS .

Once an individual is identified as having JPS , a thorough history and physical examination should be performed, with particular emphasis on the present or absence of rectal bleeding/prolapse, anemia, constipation, obstruction, diarrhea, and abdominal discomfort. Symptomatic individuals should have upper and lower endoscopies, while asymptomatic individuals should have endoscopic screenings beginning at age 15, with all polyps removed. If the initial endoscopic examinations reveal no polyps, they should be repeated in 3 years. If polyps are found, endoscopies should be repeated every year, with all polyps removed. Once no more polyps are identified, endoscopic screening should performed once every 3 years. Individuals related to the proband without an identified mutation should been screened at age 15 and if no polyps are identified screening should be done 10 years to age 45 and then as the general population if no polyps are found [252].

Children with severe bleeding, iron-deficiency anemia, hypoproteinemia, failure to thrive, and/or a non-reducible intussusception should receive an ileorectal anastomosis [253]. Individuals without these severe symptoms should receive an ileorectal anastomosis by age 20–25, when the cancer risk becomes greater than risks associated with surgery. Individuals receiving an ileorectal anastomosis should undergo flexible sigmoidoscopy every 1–3 years for polyp formation in the rectal remnant [254]. Last, individuals with generalized JPS should undergo upper endoscopy every 3 years or annually if polyps are present, beginning at age 15, due to the increased upper gastrointestinal malignancies seen in generalized JPS [241, 242, 254]. Gastric polyps tend to be diffuse and difficult to endoscopically remove, hence a total of subtotal gastrectomy should be performed where gastric outlet obstruction, adenomatous changes, or adenocarcinoma are identified [254].

PJS is inherited in an autosomal dominant pattern occurring in ~1 in 60,000 and is characterized by pigmented mucocutaneous lesions, especially on the lips, buccal mucosa, and digits, as well as hamartomatous polyps mainly seen in the small intestine. The first detailed description of PJS was given by Peutz in 1921, who described a Dutch family with familial gastrointestinal polyposis, mucocutaneous lesions, and increased cancer [255]. Individuals with PJS have a ~15-fold increased risk for developing intestinal malignancies compared to the general population, with 50 % of PJS individuals dying of cancer by age 57 [255]. The increased malignancies include small intestinal, colonic, esophageal, gastric, breast, lung, pancreatic, uterine, ovarian, testicular, and cervical cancers. Other manifestations of PJS include intussusception, abdominal pain, intestinal polyp prolapse, and hematochezia [255–258]. Interestingly, women with PJS have a high incidence of cervical adenoma malignum and ovarian sex cord tumors with annular tubules, two very rare, well-differentiated neoplasms [255–258].

PJS polyps are characterized by bundles of smooth muscle between non-neoplastic epithelium, typically having an elongated arborizing, Christmas tree-like pattern. Endoscopically PJS polyps do not have specific features and their identification requires histological examination [255–258]. The diagnosis of PJS is based on clinical findings, although molecular-genetic testing is commonly performed for prenatal diagnosis, mutation-specific kindred testing, and clinical diagnosis confirmation. The syndrome is present when any of these conditions are met: (1) presenting with 2 or more PJS polyps, (2) having one polyp and typical PJS pigmented lesions, and (3) having one PJS polyp and a family history of PJS [259].

PJS is caused by mutation of LKB1 or STK11 (serine/threonine kinase 11), found at 19p13.3. STK11 consists of 10 exons spanning 23 kb encoding a nuclear 48.6-kDa protein. STK11 is likely regulated by phosphorylation by cAMP-dependent protein kinase A and prenylation [260]. It functions as a tumor suppressor protein, with roles in the G₁ checkpoint, p53-initiated apoptosis, and regulation of Wnt-β catenin-, vascular endothelial growth factor-, and mammalian target of rapamycin (mTOR) signaling pathways [261–265]. More than 140 different STK11 mutations have been identified in individuals with PJS, with many resulting in inactive truncated kinases. Germline STK11 mutations have been found in 50–70 % of PJS kindreds and 12–50 % of sporadic PJS cases [266–269]. The reason for the relatively low detection rate is likely due to genes other than STK11 being mutated in PJS and STK11 mutations being missed due to the limitations of specific molecular tests [270, 271].

STK11 mutation testing has been performed with a variety of techniques, including DNA sequencing (often limited to exons), CSGE, SSCP, Southern blotting, real-time PCR, quantitative fluorescent PCR, denaturing gradient gel electrophoresis, long range PCR, and MLPA [272]. Volikos et al. [273] analyzed STK11 mutations via MLPA and found 39 % (11 of 28 individuals with PJS) had large STK11 deletions ranging from loss of one exon to whole gene deletions. Taken together, ~59 % of STK11 mutations (range 10–100 %) are detectable by standard methods such as SSCP, CSGE, and DNA sequencing [269, 273–280]. Another 10 % of PJS cases are detectable by RNA based methods only and ~30 % of cases require MLPA testing to detect larger deletions [273, 274, 279, 280]. Thus, current molecular diagnostics allows 80–85 % STK11 mutation identification in PJS. Currently, most clinical testing employs sequencing of the 10 STK11 exons and/or MLPA, with the later technique most being effective at detecting larger deletions [272, 274].

Due to their high risk for carcinomas, all individuals with PJS must be enrolled in lifelong cancer screening programs. While many individuals with PJS develop clinically apparent gastrointestinal lesions as early as their second decade, there are examples of malignancies in young patients and children [281, 282]. Presently there are no organization-approved guidelines for cancer surveillance in PJS patients and different institutions use different protocols [282]. The protocol used at the Johns Hopkins Hospital includes an initial baseline upper endoscopy and small bowel series at 8 years. If polyps are detected, upper endoscopy and small bowel series should be performed every 2–3 years. Upper endoscopy and small bowel series should be done annually from age 18. Colonic endoscopy should be performed every 2–3 years from age 18. At age 25–30, pancreatic endoscopic ultrasound should be done every 1–2 years. CT scanning and CA19-9 levels are offered as options. For women, a pelvic exam with a Pap smear should be performed annually from age 25. At this age, an annual transvaginal ultrasound and serum CA-125 should be done to screen for ovarian cancer. Last, at age 18, monthly breast exams should be done and at 25 semiannual clinical breast exams and annual mammography should be performed. Prophylactic mastectomy should be discussed with the patient. Males should have testicular exams and/or ultrasound of the testicles every 2 years until age 12 [283]. Recently, double balloon enteroscopy and intra-operative enteroscopy have been used to both facilitate better removal of upper gastrointestinal polyps and improve small bowel surveillance [282].

The macrolide compound rapamycin is an effective inhibitor of mammalian target of rapamycin, a cell growth-promoting kinase hyperactivated in PJS [265]. Rapamycin decreases polyp formation in PJS lkb1 +/− mice [284]. Currently, there are several clinical trials examining the effect of rapamycin on polyp formation in PJS patients [282]. Similarly, PJS lkb1 +/− mice treated with the COX-2 inhibitor celecoxib show reduced new polyp formation and a decreased size of preexisting polyps [285]. Two of 6 PJS patients given 200 mg celecoxib twice daily for 6 months showed a significant reduction in gastric polyps, suggesting that COX-2 inhibition may have some value in treating PJS [285].

Irinotecan is a topoisomerase 1 inhibitor approved by the FDA in 1994 [14, 15]. Irinotecan stabilizes the complex between topoisomerase 1 and DNA, causing DNA breaks, DNA synthesis inhibition, and apoptosis [289]. Irinotecan is itself relatively inactive, but is converted to the 100–1,000-fold more active metabolite 7-ethyl-10-hydroxy-camptothecin (SN-38) by tissue and serum carboxylesterase-converting enzymes [286, 290]. SN-38 is glucuronidated and inactivated by hepatic UGT1A1 [286]. The dose-limiting toxicities of irinotecan are severe diarrhea and myelosuppression [287, 288, 291]. The relative activity of UGT1A1 determines the effective in vivo concentration and exposure time to SN-38 [292]. Thus, decreased UGT1A1 expression may lead to a toxic reaction secondary to increased SN-38 levels and longer exposure time [287, 288, 291, 292].

UGT1A1 is located at 2p37, spans ~160 kb, and has at least 9 promoters and first exons that can be spliced to four common expressed exons, resulting in 9 different UGT1A1 enzymes, with the most common “wild-type” gene/protein designated UGT1A1*1 [293, 294]. More than 60 UGT1A1 coding and promoter variations have been identified, which lessen UGT1A1 enzyme activity [295]. The most common polymorphism, UGT1A1*28 , contains 7, rather than 6, TA promoter repeats resulting in 70 % lower UGT1A1 expression [296, 297]. TA₇ homozygotes are relatively slow to glucuronidate/inactivate SN-38, TA₆ homozygotes are more efficient, and TA₇/TA₆ heterozygotes show intermediate activity. As expected, individuals with TA₇ homozygotes are more likely to have severe diarrhea and neutropenia following irinotecan therapy, compared to TA₆/TA₆ and TA₇/TA₆ individuals [298]. Interestingly, individuals homozygous for TA₅ and TA₈ promoter repeats occur in African populations, with TA₈ associated with Gilbert syndrome and lower in vitro glucuronidation rates [299].

Other polymorphisms, such as variants at codons 71, 486, 229, and 233, have amino acid substitutions that significantly reduce enzymatic activity and SN-38 clearance. For example, a T → G transversion at codon 486 (UGT1A1 *7) substitutes a tyrosine moiety for an aspartic acid, resulting in 93 % reduce enzyme activity [300, 301]. Following irinotecan therapy, UGT1A1*28 homozygotes show an increased incidence of irinotecan toxicity compared to UGT1A1*1 homozygotes, with other variants such as G⁷¹R (UGT1A1*6) predictive of severe irinotecan-induced neutropenia and poor chemotherapeutic response in Asian populations [302–304]. Many other polymorphisms exist that alter irinotecan therapy responses, often showing population-related distribution [303, 304]. Not surprisingly, individuals heterozygous for two low-activity UGT1A1 alleles often exhibit irinotecan toxicity. For example, Takane et al. [305] reported severe irinotecan toxicity in a UGT1A1*6/*28 patient.

UGT1A1 allelic variations have been detected through a variety of techniques, including analysis of single base extension of PCR-amplified products, DNA sequencing, HRM, radioactive PCR, denaturing high performance liquid chromatography, and Invader technology. Currently, Invader technology is FDA-approved for UGT1A1 molecular testing and is commonly used [297, 302, 306–309].

Invader technology typically distinguishes between two different alleles within one non-PCR-based reaction. The reaction involves two steps. First, two allele-specific oligonucelotides, termed Invader oligonucleotides, bind to an unlabeled primary probe. The two Invader oligonucleotides hybridize in tandem with the primary probe forming a specific overlap structure. The 5′-end of the primary probe contains a non-complementary sequence that does not hybridize with either Invader oligonucleotide. The 3′-end of the Invader oligonucleotides are designed to overlap the primary probe such that one of the Invader oligonucleotides will form a complementary base pair with the primary probe. This bp (or pairs for the *28 TA₇) correspond to the wild-type and polymorphism being tested. The complementary base pairing does occur, the resulting structure is recognized and cleaved by the Cleavace® enzyme (Flap endonuclease), releasing the 5′ portion of the primary probe. Should the 3′ portion of the Invader oligonucleotide not be complementary, the Cleavace® enzyme will not cut the structure.

In the second reaction, a second probe binds to the cleaved 5′primary probe sequence, forming a similar structure that is cleaved by the Cleavace® enzyme. The second probe is designed such that the second cleaving reaction separates a fluorophore from a quencher, creating an easily measured signal. This system can incorporate two primary probes that will eventually cleave either one of two secondary probes, each labeled with a spectrally distinct fluorophore corresponding to a specific allele. The ratio of one fluorophore to the other allows allele identification.

The accuracy of assay comes from the extreme specificity of the Cleavace® enzyme and the sequence specificity of the oligonucleotide hybridization. Purified genomic DNA may be used without PCR amplification and the system has the advantage of being completely closed once the assay has begun. Invader technology is a medium through-put, relatively low-cost technology that commonly allows up to 96 patient samples and controls to be simultaneously interrogated [308, 309].

Despite the large number of different techniques employed in UGT1A1 genotyping, the analytic sensitivity was 100 % for the UGT1A1*28 and 98–100 % for the *6 allele, where the referent test was DNA sequencing [310]. The reproducibility of Invader technology was 98.8 % [309]. Additionally, estimates of the first run failure UGT1A1*1 and UGT1A1*28 analysis was 5.0 % for sequencing, 1.7 % for PCR/capillary electrophoresis, and 6.7 % for Invader® technology, with most discrepancies solved by repeat testing [309].

The risk of a toxic irinotecan reaction is roughly 50, 12.5, and 0 % for *28/*28, *28/*1, and *1/*1 individuals, respectively [297]. However, there is significant overlap between these groups, and despite a large number of studies, the association between the *28 homozygotic genotype and irinotecan toxicity is not statistically significant [309]. Additionally, at standard irinotecan doses, the tumor response rate is 41 % for wild-type, 45 % for *28 heterozygotes, and 70 % for *28 heterozygotes [309]. Thus the wild-type individuals may be under dosed, while the *28 heterozygotes show a better response. Based on this, The Evaluation of Genomic Applications in Practice and Prevention Working Group found insufficient evidence to recommend for or against UGT1A1 genotyping guiding irinotecan treatment protocols [309].

Recently however, one Phase I study was performed on 27 patients with mCRC, examining irinotecan toxicity and the *28 polymorphism: 18 patients were *1/*1 and 9 were *1/*28. A maximum tolerated dose was not observed up to a 150 mg/m² dose in the *1/*1 patients, while the *1/*28 patients showed toxicity at a 100 mg/m² irinotecan dose [310]. The study concluded that *1/*1 individuals should receive 150 mg/m² dose irinotecan, while *1/*28 individuals should receive 70 mg/m². The study was limited to 27 patients and included no *28/*28 individuals. Most likely, further studies will be performed and may demonstrate clinical utility for UGT1A1 genotyping [310].

Murff et al. [311] found that only 17 % of individuals who met the criteria for early-onset breast cancer testing were actually referred for genetic testing. Presently, many individuals who would benefit from molecular testing do not receive it. As the clinical utility of molecular testing becomes more apparent and testing costs fall, this diagnostic modality will be increasingly employed.

Presently, UGT1A1 genotyping to guide irinotecan use is not routinely performed, although an initial study suggests that it might have use in determining the optimal irinotecan dose [310]. Additionally, Bedoya et al. [312] found that Dukes stage C and D patients were significantly more likely (P = 0.05) to have a position 1,359 G → A nucleotide change in the cannabinoid receptor-1. They also had a shorter survival time and increased lymph node metastasis. While not currently a part of molecular diagnostics, increased understanding of mutations/polymorphisms in other genes effecting CRC genesis and progression, such as the cannabinoid receptor-1, will likely become important in the molecular diagnostics of this malignancy [14–17, 311].

Several molecular techniques, while not often employed in clinical testing, have proven to be very useful in the analyses of CRC and are very likely to significantly contribute to the molecular diagnostics of CRC in the near future.

Microarray. Micoarray consists of a solid support, often a glass slide, to which DNA probes of known sequence are fixed in an orderly arrangement. The slides are hybridized under high-stringency conditions with labeled nucleic acids (often cDNAs derived from RNAs), extensively washed, and the relative amount of probe-bound target sequence is measured via label detection. The number of gene/nucleic acid sequences that can be simultaneously measured can range from 10 to 1 million. Comparisons of normal colonic mucosa and CRC have shown a large number of gene products over-or under-expressed in CRC compared to normal colonic tissue [313]. Microarray has allowed the development of specific gene expression panels with diagnostic (Pathwork test for tissue of origin) or prognostic (OncotypeDX) abilities. For CRC, specific gene products have proven useful in predicting clinical behavior. For example, osteopontin RNA expression is a well-documented biomarker of CRC progression [314].

Proteomics . Proteomics involves the analysis and identification of proteins produced by cells in normal and pathologic conditions. Often these analyses examine protein modifications, such as phosphorylation and protein-protein interactions. The specific techniques used ranges from antibody-based protein chips resembling microarray techniques, to yeast two-hybrid systems to identify protein-protein interactions, high throughput crystallography screens for structural analyses, two-dimensional electrophoresis, and mass spectrometry. Although preliminary, these analyses have proven very useful in identifying proteins, whose expression correlates with chemotherapy response, metastatic potential, and degree of clinical aggressiveness. With time, specific proeomic profiles can be developed to predict the clinical behavior of specific CRC cases [313].