Author reply

Abstract

We thank Schulmann et al. for the valuable comments related to our article.1 We have rechecked the microsatellite instability (MSI) status and performed additional immunostainings in our material and would like to comment on the questions raised. The objective of our study was to evaluate the frequency of defective mismatch-repair (MMR) in a population-based series of small bowel carcinomas in the Southern Sweden health care region, 1989–1999 (Series I), and in patients who were diagnosed before age 60 (Series II). Because the paraffin-embedded tumor blocks were collected from different pathology departments, and the majority of the patients were diseased, data on family history of cancer were not available. For practical reasons, only tumor-containing blocks were available; therefore, we chose to use mononucleotide markers for the MSI analysis, because these markers can be interpreted without the need for matching normal tissue. However, we agree that an MSI classification based on the instability of BAT40 is suboptimal, although normal allelic size variations as large as those classified as unstable in our material are rare.2 Hence, we cannot rule out the possibility that these two tumors represent false-positive results. Because data on MSI and immunostaining in carcinoma of the small intestine are scarce, it is not known which markers detect MSI with the greatest sensitivity in this tumor type. The sensitivity of the MSI markers depends on the tumor type, as demonstrated, e.g., for colorectal carcinoma and endometrial carcinoma. Furthermore, even if all 4 tumors with low MSI (MSI-L; classified based on data from 3 MSI markers, of which only one was positive) in the population-based series had been excluded, only 7 of 11 tumors (65%) with high MSI (MSI-H) showed loss of MLH1 or MSH2. We do not believe that the MSI-H tumors with retained immunostaining represented false negative MLH1 staining; immunostaining was retained in the stromal components and in the tumor-infiltrating lymphocytes of all these tumors. We routinely perform MSI and MMR protein immunostaining to identify hereditary nonpolyposis colorectal cancer in patients with colorectal carcinoma and have previously experienced a high degree of correlation between the MSI and MMR protein immunostaining. The two tumors with microsatellite stability (MSS) that showed a loss of MLH1 staining likely represent false negative MSI tumors, and although it is a rare finding, loss of MMR protein expression in MSS tumors has been described by several authors and often has been associated with suboptimal archival tumor tissue.3 To characterize the genetic defect behind the MSI tumors with retained expression of MLH1 and MSH2, we have immunostained the 23 evaluable tumors (described in Table 1 of our recent article1) for the MMR protein MSH6. In the population-based series (Series I), one tumor (X42) showed loss of MSH6 staining, and one tumor (X115) among the young patients (Series II) also showed loss of MSH6 alone. In addition, loss of MSH6 was (as expected) observed in three tumors (X123, X134, and X139) with loss of MSH2 staining. Hence, after inclusion of MSH6 staining, loss of MMR protein staining was found in 7 of 15 MSI tumors in the population-based series and in 8 of 10 tumors among individuals younger than age 60 years. In summary, adding immunostaining for MSH6 indeed may explain an additional two discordant (MSI and retained immunostaining for MLH1 and MSH2) tumors. However, in the whole material retained staining for the MMR proteins MLH1, MSH2, and MSH6 still was present in 10 of 21 MSI tumors. Therefore, these findings suggest that other mechanisms may cause a subset of the MMR-defective carcinomas of the small bowel.