|Home | About | Journals | Submit | Contact Us | Français|
Though the presence of microsatellite instability (MSI) in patients with colorectal cancer (CRC) may have implications for prognosis, therapy, and family counseling, MSI prevalence is not well described among individuals of Hispanic origin in the United States (US) with CRC.
We conducted a retrospective cohort study employing a hospital-based tumor registry to identify individuals of Hispanic origin diagnosed with CRC. Clinical data and tumor samples were retrieved. Molecular analyses included testing for MSI using a panel of 5 mononucleotide markers (BAT25, BAT26, NR21, NR24 and NR27) in a pentaplex polymerase chain reaction assay, as well as immunohistochemistry for the MMR proteins MLH1, MSH2, MSH6, and PMHS2 on representative tissue.
111 individuals of Hispanic origin with CRC were identified. 41.1% were women, median age was 57 years (IQR:47.1–63.5). 11 (9.8%, 95%CI:4.2–15.6) had MSI CRC, while 14 (14.6%, 95%CI:7.3–21.8) had CRC with≥1 MMR protein abnormality. 10 of 11 individuals with MSI had clinical or molecular characteristics suspicious for Lynch syndrome such as abnormal expression of MSH2 and/or MSH6 (n=7) or age<50 at diagnosis (n=7).
The prevalence of MSI CRC among Hispanic individuals may be similar to other races and ethnicities, but clinical-pathological characteristics, including age at diagnosis and pattern of abnormal MMR protein expression, suggests that sporadic MSI CRC may be less common in individuals of Hispanic origin, and that much MSI observed in this situation may be attributable to Lynch syndrome. Further exploration of the causes of disparate presentations of CRC by ethnicity and race is warranted.
Approximately 7 to 14% of colorectal cancer (CRC) is associated with microsatellite instability (MSI) 1–4. MSI is due either to acquired (sporadic MSI) or inherited (Lynch Syndrome) defects in DNA mismatch repair mechanisms. Knowledge of CRC MSI may have clinical implications for prognosis, therapy, and family counseling. Individuals with MSI CRC may have more favorable survival compared to those with microsatellite stable (MSS) CRC, and require more tailored clinical care 1, 5–8. The prevalence of MSI CRC has been described among Whites and African-Americans in the United States (US) 2, 4, 9. However, the prevalence of MSI among individuals of Hispanic origin in the United States is not well described. CRC is the third leading cause of cancer death among Hispanic men and women, accounting for 16.1 deaths/100,000 men, and 10.7 deaths/100,000 women--rates somewhat lower than that reported for non-Hispanic men and women10, 11. Thus, given the significant burden of CRC among Hispanics the scarcity of data on MSI among Hispanics with CRC, and the emerging importance of MSI status on clinical management of CRC, the prevalence of MSI among individuals of Hispanic origin requires further characterization1, 5–8.. Here, we hypothesized that the prevalence of MSI and mismatch repair protein (MMR) protein abnormalities among Hispanic individuals with CRC may play a role in the clinical presentation of patients with this ethnicity. Our primary aim was therefore to study the prevalence and clinical presentation of MSI CRC in a series of individuals of Hispanic origin diagnosed and/or treated for CRC at a large, urban US hospital.
Study subjects were identified by using a hospital-based, American College of Surgeons certified tumor registry at Parkland Memorial Hospital--a large, urban county hospital serving residents of Dallas County, Texas. Study subjects included were of Hispanic origin, with a primary adenocarcinoma of the colon or rectum diagnosed and/or treated at Parkland Memorial Hospital at any time between 1997 and 2007, and sufficient archived, paraffin embedded tissue for MSI analysis. Hispanic ethnicity was designated based on ethnicity recorded in the demographic section of the medical record through the course of usual medical care. It is our observation that at Parkland, Hispanic ethnicity is generally assigned to individuals whose primary language is Spanish, or whose family origin is from Mexico, or any country in Central America, South America, or the Caribbean where the primary language is Spanish. Subjects were excluded if sufficient tissue for MSI analysis was not available, if primary cancer was not adenocarcinoma of the colon, if cancer was recurrent rather than diagnosed at first presentation, or for presence of a known genetic cancer syndrome other than Lynch syndrome. Clinical data, including age, sex, SEER (Surveillance, Epidemiology, End Results) summary stage, date/survival status at last contact, and right-sided CRC location (defined as proximal to the descending colon), as well as archived, paraffin-embedded tumor samples were retrieved. Pedigrees were not routinely generated as part of usual clinical care, thus detailed family history data were not available. Histology of the primary tumor was re-reviewed by a dedicated pathologist for the presence of characteristics consistent with MSI, including histological type, degree of differentiation, tumor-infiltrating lymphocytes, and a Crohn’s like lymphocytic reaction 12. The study was approved by the University of Texas Southwestern Medical Center Institutional Review Board.
After microdissection and capture of representative CRC tissue, MSI analyses were performed using a panel of 5 mononucleotide markers (BAT25, BAT26, NR21, NR24 and NR27) in a pentaplex PCR assay as described previously 13. Mutations of 2 or more markers as positive were considered evidence of MSI.
All immunostaining was performed at room temperature and carried out on the Dako Autostainer (DakoCytomation-Carpinteria, CA). Reagents were used as supplied in the Mach 4™ Universal HRP-Polymer Kit (BioCare Medical, Walnut Creek, CA). BioCare Diva, ph 6.6 was used. Optimum primary antibody dilutions were predetermined using known positive control tissues. A known positive control section was included in each run to ensure proper staining. Antibody dilutions were as follows: MLH1 - 1:600 (G168-728, BD Pharmigen); MLH2 - 1:800 (FE11, BD Pharmigen); MSH6 - 1:800 (GTBP, BD Transduction); PMS2 - 1:150 (A16-4, BD Sciences). Paraffin sections (5μm) were cut on a rotary microtome, mounted on positively charged glass slides (Superfrost, Fisher), and baked one hour at 65 degrees Celsius. Sections were then deparaffinized in xylene followed by ethanol and then placed in 200ml Diva, ph 6.6. The buffer was brought to 95 degrees Celsius for 40 min. Sections were allowed to cool in buffer for 5 min, and then rinsed in deionized water. The sections were then loaded onto the Dako Autostainer. Sections were first quenched with 3% H2O2 for 5 min, rinsed in buffer, and incubated with primary antibody at the concentrations highlighted above. Following wash of the primary antibody, sections were incubated for 15 min in BioCare Mach 4™ Mouse Probe rinsed in buffer and then incubated for 15 min in MR HRP-Polymer. After rinsing in buffer, the sections were then incubated in a freshly prepared mixture of diaminobenzidine (DAB) and buffer substrate solution for 5 min. Sections were then counterstained with hematoxylin and blued in Richard Allen Bluing Reagent ™, dehydrated in a graded series of ethanols and xylene, and coverslipped. Slide analysis was performed on Automated Cellular Imaging (Dako, CA) by choosing ten random tumor fields to determine percentage of positive tumor cells and intensity of staining. For primary analyses, staining expression of <10% percent positive cells or staining intensity of <2 for any proteins was considered evidence of abnormal (decreased) protein expression.
Descriptive statistical calculations, including means (and medians when data were non-normally distributed) were performed. Chi-squared tests (or Fischer’s exact tests for comparisons with fewer than 5 observations for cells) were employed for comparisons of proportion. Survival analyses, comparing individuals with MSI and no evidence of MSI (MSS) were performed, censoring individuals at time of last follow up or death. Though family history data were not available for analyses, we assessed whether cases with MSI and/or abnormal expression of an MMR protein were suspicious for Lynch syndrome based on clinical or molecular characteristics more likely to be associated with Lynch syndrome rather than sporadic MSI CRC, such as age of presentation younger than 50, presence of abnormal MSH2 and/or MSH6 expression, and synchronous CRC. All analyses were performed with STATA 9.0, College Station, TX.
From 933 individuals of all races and ethnicities with a colon or rectal neoplasm diagnosed and/or treated at Parkland Memorial Hospital between 1997 and 2007, we identified 161 potential study subjects of Hispanic origin (Figure 1). A total of 50 subjects were excluded for the reasons highlighted in Figure 1; the most common reasons for exclusion were presence of a non-primary adenocarcinoma (n=24) such as anal carcinoma, and absence of tissue for analysis (n=22), usually due to pathologic diagnosis at another facility. Thus, 111 individuals with primary adenocarcinoma of the colon or rectum and sufficient tissue for MSI analyses were included for analysis. The median age of the cohort was 57 years and 41.1% of subjects were women (n=46, Table 1). Thirty-two percent were younger than age 50 at diagnosis (n=36). Nearly 68% presented with SEER summary stage regional or distant CRC (n=75). Over half were alive at last follow up. Median follow up time ranged from over 9 months for those expired at last follow up, to nearly 54 months for those alive at last follow up.
Prevalence of MSI in the cohort was 9.9% (n=11, 95% CI: 4.2–15.6) (Figure 1). Figures 2a and 2b depict examples of subjects with abnormal expression of MMR proteins. Prevalence of abnormal expression of at least one MMR protein among all 111 individuals in the cohort was 14.6% (n=14, 95%CI: 7.3–21.8). All individuals with MSI CRC had abnormal MMR protein expression by IHC, while 3.5% of individuals with MSS CRC had abnormal IHC (Table 2). Fifteen individuals had incomplete analysis for all 4 MMR proteins due to technical limitations. Histology suggestive of MSI was present in 36.9% of the cohort (n=41, 95% CI: 27.8–46.1), and 5 of the 11 MSI CRC cases had histology suspicious for MSI. MSI CRC was associated with age<50 and right-sided tumor location within the colon, but not female gender or MSI-type histology, though point estimates for odds ratios were consistent with an association with these factors as well (Table 3). Survival appeared to be more favorable among individuals with MSI, though this did not reach statistical significance (p=0.11, log-rank test for equality of survivor functions, Figure 3).
A summary of clinical and molecular characteristics in specific subjects with MSI is shown in Table 4. Seven of 11 were younger than age 50 at diagnosis, 7 of 11 had abnormal expression of MSH2 and/or MSH6, and 4 of 11 individuals were younger than age 50 with abnormal expression of MSH2 and/or MSH6. Two individuals had synchronous CRC, though one of these individuals was age 81. Characteristics of the 3 subjects with normal MSI testing and abnormal MMR protein expression are summarized in Table 5. Detailed review suggested that MSI testing may have been falsely negative due to the presence of a very small tumor in one case; no clear reason for false negative MSI testing (or false positive IHC) could be found for the other two cases.
Our results demonstrate that the prevalence of MSI among individuals of Hispanic origin with CRC diagnosed at a large, urban, US County hospital is approximately 10%. Data from a comparable study from Puerto Rico are consistent with our findings. De Jesus-Monge et al. evaluated prevalence of abnormal IHC for the MMR proteins MLH1 and MSH2 among 164 individuals in Puerto Rico14. The prevalence of abnormal expression of either MLH1 or MSH2 was 4.3%, but immunohistochemistry for PM2 and MS6 was not performed, and MSI testing was restricted to those with abnormal IHC, perhaps leading to an underestimation of CRC attributable to MMR protein dysfunction. The single patient with abnormal IHC for MLH1 had MSI but normal testing for a BRAFV600E mutation. Consistent with our study, the pattern of molecular abnormalities suggested that the prevalence of sporadic MSI among Hispanics was low, and that the abnormalities seen were suspicious for underlying Lynch syndrome. Similar to our study, germline testing for confirmation was precluded by the retrospective nature of the study. Overall, in considering other studies the estimated prevalence of MSI CRC found in our study is similar to the 7 to 13% prevalence reported in population-studies from Italy, Spain, and other parts of the US1–3, 15, 16, but lower than the 20% prevalence reported in a previous large US study17, suggesting that our reported prevalence is similar to non-Hispanic populations. Interestingly, 10 of 11 subjects with MSI had clinical and/or molecular characteristics that raise suspicion for Lynch syndrome, suggesting that sporadic MSI may be uncommon among Hispanics. The one exception was an individual most likely to have sporadic MSI secondary to hypermethylation associated with age and type of MMR abnormality, as he was age 81 at diagnosis with abnormal MLH1 and PMS2 expression18. If true, a low prevalence of CRC attributable to sporadic MSI among Hispanics may be explained by a lower frequency of exposure (or resistance to) environmental carcinogens, or a lower frequency of a genetic predisposition to develop the alterations that result in sporadic MSI CRC in the non-Hispanic population.
Our findings may have implications for clinical care and future research. While adaptation of universal screening for MSI and Lynch syndrome for all individuals with CRC is under debate on a national level19, 20, on a local level, if the prevalence of MSI CRC is indeed 10%, then universal screening may be worthwhile, as identification of individuals with MSI may result in more optimized surgical care, planning of post-resection surveillance, and family risk counseling, and help reduce disparities between Hispanics and Whites in CRC survival21, 22. Indeed, while survival was not statistically significantly different between Hispanic individuals with MSI and MSS CRC in our study, the curves are clearly divergent and approach significance (Figure 3). We postulate that with a larger sample size, the favorable survival observed for individuals with MSI as compared to those with MSS CRC in other studies may have been observed, extending the potential importance of MSI testing for prognosis and management of Hispanic individuals with CRC. From a research perspective, if the prevalence of sporadic MSI CRC is indeed low, comparative studies of environmental and lifestyle factors between Hispanic and non-Hispanic individuals may offer insight into exposures that may be important to the development of sporadic MSI CRC. For example, our findings may reflect lower exposure to “Western” lifestyle factors associated with MSI CRC, such as tobacco 2, 4, 23, 24, that may be less prevalent but increasing over time and across generations among Hispanics in the United States25–27. Further study of these factors and others is warranted, particularly if the prevalence of sporadic MSI is low.
We recognize several potential limitations to our findings and conclusions. First, due to the retrospective nature of our study, detailed family history and pedigree data were not available for analyses, and blood samples for sequencing to identify germline mutations in MMR genes were not obtainable. We can therefore only speculate as to whether the MSI cases identified were a result of hereditary or acquired genetic changes. It is well known that while the Bethesda criteria and presence of abnormal MMR expression are associated with increased likelihood of Lynch syndrome, the specificity of these characteristics are suboptimal, and thus insufficient to make definitive conclusions regarding the presence/absence of Lynch syndrome among our MSI cases 16. Second, neither the country of origin, nor birth generation relative to immigration to the US was available. It has been observed that CRC incidence is highly modified by degree of industrialization and country of origin presumably due to country specific differences in lifestyle and environmental risk factors 28. If risk for MSI CRC differs by country of origin based up on diet, lifestyle, or hereditary risk, it is possible that our findings are not generalizeable to all individuals of Hispanic origin living in the US. Third, Parkland Memorial Hospital serves a disproportionate number of uninsured patients in the Dallas area, perhaps resulting in selection bias towards a younger population with CRC. Sporadic MSI is generally associated with an age at presentation above 65, when the likelihood of access to insurance is higher in the US, while Lynch syndrome related MSI CRC is more common among younger individuals who may be more likely to be uninsured and present to Parkland if residing in Dallas County 29. This may have resulted in under-estimation of the prevalence of sporadic MSI (and perhaps overall MSI), and over-estimation of the prevalence of suspected hereditary MSI CRC. Fourth, identification of individuals of Hispanic ethnicity was based on retrospective review of demographic data collected through the course of usual clinical care, perhaps leading to decreased sensitivity or specificity for identifying individuals of Hispanic origin with CRC through our tumor registry. Finally, direct comparisons to non-Hispanic whites and African-Americans were not performed; therefore, prevalence relative to these groups could not be assessed.
The strengths of our study include the use of a large, comprehensive, American College of Surgeons certified cancer registry to identify cases that resulted in a large sample size, analysis of MSI in all cases, analysis for abnormal MMR expression in nearly all cases, and abstraction of detailed clinical information from clinical and cancer registry records. Extensive efforts were taken to include every potential case, without bias towards selection from the potential study pool.
In conclusion, the prevalence of MSI CRC among Hispanic individuals may be similar to other races and ethnicities, but clinical and molecular characteristics, including age at diagnosis and type of abnormal MMR protein expression, suggests that sporadic MSI may be less common in individuals of Hispanic ethnicity, and that much MSI observed in this situation may be attributable to Lynch syndrome. Further exploration of the causes of disparate disease presentations of CRC by ethnicity and race, including characterization of sporadic and hereditary MSI, is warranted.
Sources of Support: This work was supported by an American Cancer Society Institutional Research Grant to UT Southwestern Medical Center, ACS-IRG-02-196, as well as National Institutes of Health grant number 1 KL2 RR024983-01, titled, “North and Central Texas Clinical and Translational Science Initiative” (Milton Packer, M.D., PI) from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH) and NIH Roadmap for Medical Research, and its contents are solely the responsibility of the authors and do not necessarily represent the official view of the NCRR or NIH. Information on NCRR is available at http://www.ncrr.nih.gov/. Information on Re-engineering the Clinical Research Enterprise can be obtained from http://nihroadmap.nih.gov/clinicalresearch/overview-translational.asp. Additional support was provided by National Institutes of Health grant number R01 CA72851-15, titled “Familial and Early Onset Colorectal Cancer,” (C. Richard Boland, PI).
We are grateful to Darnetta Miles and staff at the Parkland Memorial Hospital Clinical Pathology Service, as well as staff at the Parkland Memorial Hospital Cancer Registry for assistance with obtaining study data and materials.
Financial Disclosures: None applicable.