Histomorphology and Molecular Genetics in Different Intra-Tumoral Zones and Matched Metastatic Lymph Nodes of Colorectal Cancer with Heterogenous Mismatch Repair Status


 Objective: To better understand the clinicopathological characteristics and molecular alterations in different intra-tumoral components of colorectal cancer (CRC) with heterogeneity of mismatch repair (MMR) protein expression and microsatellite instability (MSI) status.Methods: We identified 4 cases of CRC with heterogenous MMR protein expression and analyzed the histopathological features, MSI status and other molecular alterations in separately microdissected intra-tumoral zones and lymph node metastases by polymerase chain reaction (PCR) -based MSI testing, MLH1 promoter methylation and targeted next-generation sequencing (NGS).Results: Microsatellite instability-high (MSI-H) was identified in the MLH1/PMS2 deficient zones in Case 1-3, and in the MSH2/MSH6 deficient zone in Case 4, while MSS was in all the intra-tumoral zones and metastatic lymph nodes with proficient MMR (pMMR). Furthermore, heterogeneity of MLH1 promoter methylation and/or other common driving gene mutations of CRC, such as KRAS, PIK3CA and so on, was identified in all the 4 CRCs. In addition, 75% (3/4) of cases showed heterogeneity of histomorphology in intra-tumoral components and metastatic lymph nodes (Case 1, 2, 4), and all the corresponding metastatic lymph nodes were moderate differentiation with MSS/pMMR (Case 2, 3). Conclusions: The heterogeneous MSI status is highly correlated with histomorphological heterogeneity, which is also an important clue for the heterogeneity of drive gene mutations in CRC. These results suggest that it is essential to detect MMR protein expression and other gene mutations in metastases before treatment, especially for the CRCs with heterogenous MMR protein expression or histomorphology.

status in paired primary and metastatic CRC from 369 patients and found that 19.6% (9/46) had an MSS metastasis in patients with an MSI-H primary tumor (inter-tumoral heterogeneity), and the discrepancy was more likely to be limited to peritoneal or ovarian metastasis [15]. Therefore, heterogeneity of MMR protein expression and MSI status in CRC is rare but exists indeed. However, clinicopathology and molecular genetics in different intra-tumoral zones and matched metastatic lymph nodes of colorectal cancer with heterogenous MSI status are still not clear.
In this study, we collected 4 CRCs with heterogenous MMR protein expression and analyzed the clinicopathological and immunohistochemical characteristics. More importantly, we determined the MSI status, MLH1 promoter methylation, and other molecular alterations in separately microdissected intratumoral components and matched metastatic lymph nodes with heterogenous MMR protein expression using PCR-based MSI testing and targeted NGS.

Patients
Four cases with heterogenous MMR protein staining were retrieved from 500 cases of CRC resections in the pathology les in the Department of Pathology, Fudan University Shanghai Cancer Center during 2018 -2020. All the cases were reviewed by two senior pathologists. Clinicopathologic information of these 4 cases was obtained from the medical records and/or discharge summary (Table1). Formalin-xed para n-embedded sections, including the primary tumor tissues, metastatic lymph nodes, and matching normal tissue sections (3-4 µm thick) were collected. The present study was approved by our institutional ethics committee. All tests of samples in this study were carried out in ISO15189 certi ed laboratories.

Microdissection And Dna Extraction
Intra-tumoral components with heterogenous MMR protein expression were separately microdissected by laser capture microdissection (LCM, Arcturus XT; Life Technologies, Mountain View, CA, USA) as previously described [19,20]. Brie y, ten 6-µm-thick histologic sections were prepared from each selected block and adhered to a 1.4-µm membrane with metal frame slides (Applied Biosystems, Foster City, Germany). After drying and dewaxing routinely, these sections were xed in 100% ice-cold ethanol for 10min, stained with hematoxylin for 1 min, and dehydrated in 100% ethanol for 30s and xylene for 5min. Target components with heterogenous MMR expression were isolated separately using an LCM system. Genomic DNA from different intra-tumoral components and corresponding metastatic lymph nodes with heterogenous MMR protein expression, as well as matching normal mucosa samples were extracted with the QIAamp DNA Mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions.

Pcr-based Msi Testing
Microsatellite instability status was evaluated by a uorescent PCR-based assay with Promega MSI Analysis System (Promega Corporating, Madison, US), which included ve mononucleotide markers (BAT-25, BAT-26, NR-21, NR-24, and MONO-27) and two pentanucleotide markers (Penta C, Penta D). The testing was conducted as previously described [21]. Brie y, Promega MSI testing was performed with a 2 ng DNA input for separately microdissected intra-tumoral components and metastatic lymph nodes with heterogenous MMR protein expression. PCR ampli cation was carried out in a GeneAmp PCR 9700 thermocycler (Applied Biosystems) and capillary electrophoresis was carried out in an ABI 3500xL automated DNA sequencer (Applied Biosystems). Data were analyzed with the GeneMapper Software Analyzer (Thermo Fisher Scienti c, Waltham, MA). Compared with the normal mucosa control tissues, tumors were de ned as MSI-high (MSI-H) when ≥ 2/5 of the mononucleotide markers were altered, MSI-low (MSI-L) with only 1 marker altered, and microsatellite stable (MSS) with stability for all the ve markers.

Methylation-speci c Pcr Analysis
Extracted DNA from different intra-tumoral components and corresponding metastatic lymph nodes with heterogenous MMR protein expression was treated with bisul te to convert unmethylated cytosines to uracil while leaving methylated cytosines unaltered. MLH1 promoter methylation status was analyzed by means of a uorescence-based, real-time methylation-speci c PCR assay using the Human MLH1 Gene Methylation Detection Kit (Gene Tech (Shanghai) Company Limited, China) according to the manufacturer's instructions. This kit analyzed the methylation of 5 CpG sites (AGAGCGGACAGCGATCTCTAACGCGCAAGCGCAT, chr3: 37, 034, 766-37,034,799, GRCh37/hg19). The PCR reactions (95℃ for 3 min, followed by 45 cycles of 95℃ for 15 sec and 60℃ for 60 sec) were performed on an ABI 7500 analyzer (Applied Biosystems, Foster City, CA). The Ct value of the sample and the external control must be less than or equal to 32. Positive results were de ned as Ct (sample) -Ct(control) ≤7. Samples were run in duplicate, including positive and negative controls.

Dna-based Next Generation Sequencing
The ColonCore panel (Burning Rock Biotech, Guangzhou, China) is designed for simultaneous detection of MSI status and mutations in 41 CRC-related genes, including KRAS, NRAS, BRAF, PIK3CA, hereditary CRC genes, MMR genes, and other genes related to carcinogenesis and tumor development. The genes included in this panel are listed in Supplementary Table 1. All the detection was carried out according to the manufacturer's instructions as previously described using a NextSeq platform (Illumina Inc., San Diego, CA) [22,23]. Brie y, DNA shearing was performed using Covaris M220 (Covaris, Inc., MA, US), followed by end repair, phosphorylation, and adaptor ligation. Fragment sizes ranging from 200 to 400 bp were selected using Agencourt AMPure beads (Beckman Coulter, CA, US) followed by hybridization with capture probes baits, hybrid selection with magnetic beads, and PCR ampli cation. Subsequently, Qubit® 3.0 and Agilent 2100 bioanalyzer (Agilent Technologies Inc., CA, US) was performed to assess the quality and size of the fragments. Indexed samples were sequenced on the Nextseq500 sequencer (Illumina, Inc., CA, US) with pair-end reads. The sequencing data were mapped to the human genome (hg19) using Burrows-Wheeler Aligner version 0.7.10. Local alignment optimization, variant calling, and annotation were performed using the Genome Analysis Toolkit version 3.2 and VarScan version 2.4.3.
The method of MSI phenotype detection in this ColonCore panel was a read-count-distribution-based approach. It took the coverage rate of a set of speci c repeat lengths as the main feature of each microsatellite locus. The determination of MSI status was not required for the matched normal tissues. If the coverage rate was less than a given threshold, the locus was classi ed as unstable. The MSI status of a sample was determined by the percentage of unstable loci in the given sample. A tumor sample was considered MSI-H if more than 40% of the marker loci were length-instable, MSS if the percentage of length-instable loci was less than 15%, or MSI-L if the percentage was between 15-40% [22].

MMR protein expression and microsatellite instability status
Four cases of CRC with heterogenous MMR protein expression were identi ed for in-depth analysis. Immunohistochemical staining was repeated to con rm the heterogenous MMR protein expression in all the 4 tumors ( Table 2). Brie y, intra-tumoral heterogenous expression affected MLH1/PMS2 in Case 1-3, MSH2/MSH6 in Case 4. Interestingly, pMMR was demonstrated in all the metastatic lymph nodes in Case 2 and Case 3.

Clinical And Histopathological Features
The clinical and histopathological features of the 4 CRCs with a heterogeneity of MMR protein expression and MSI status were summarized in Table 1. There were 2 female and 2 male patients, with ages at presentation ranging from 50 to 68 years (mean, 59.5 years). The tumor size was ranged from 1.8 to 7.5 cm. Most (3/4, 75%) of them were located in the right colon.
Histopathologically, intra-tumoral morphological heterogeneity was determined in 3 of the 4 cases, was exactly matched with the heterogeneous zones of MMR protein expression. Brie y, intra-tumoral Zone B with MSI-H/dMMR (MLH1/PMS2 loss) in Case 1 and 2 were poorly differentiated, showing a solid or dense sieve growth pattern, while intra-tumoral Zone A and matched metastatic lymph nodes with MSS/pMMR were moderate differentiation with a glandular or cribriform-like pattern (Fig. 1, C-N). On the other hand, the matching was just opposite in Case 4, intra-tumoral Zone B with MSI-H/dMMR (MSH2/MSH6 loss) was moderate differentiation and partly showed mucinous adenocarcinoma (Adca), but poor differentiation in the Zone A with MSS/pMMR ( Fig. 3, B-I). No morphological heterogeneity in primary tumor and metastatic lymph nodes was found in Case 3 (Fig. 2, C, D and E). It's worth noting that all the corresponding metastatic lymph nodes were moderate differentiation and determined to be MSS/pMMR (Fig. 1, E, H, K and N, Fig. 2, E, H, K and N). In addition, heterogeneous histomorphological components with moderate or poor differentiation in Case 1, 2, 4 were all positive for CDX2 and SATB2, and all negative for CgA and Syn (data not shown).

Mlh1 Promoter Methylation And Other Molecular Features
To further explore the possible causes of heterogenous MMR expression, MLH1 promoter methylation and targeted NGS were performed in separately microdissected intra-tumoral components and matched metastatic lymph nodes, respectively. As a result, NGS con rmed the heterogeneity of MSI status again, consistent with the above MSI-PCR results. Furthermore, signi cant differences in genetic mutations were identi ed intra-tumoral zones and matched metastatic lymph nodes with heterogenous MMR expression and MSI status, including the most common driving gene mutations related to CRCs, such as KRAS and PIK3CA, and so on. The gene mutations with pathogenic and likely pathogenic signi cance in different zones of all the 4 cases were listed in Table 3, and the gene mutations with uncertain signi cance were list in Supplementary Table 2. promoter methylation, i.e. intra-tumoral Zone B with loss of MLH1/PMS2 expression showed MLH1 methylation, whereas intra-tumoral Zone A and matched metastatic lymph nodes with pMMR did not show MLH1 methylation.

Discussion
Heterogenous MMR protein expression was a rare phenomenon in CRC. In this study, we identi ed 4 cases of CRC with heterogeneous MMR protein expression, and then analyzed the clinical and histopathological features of these cases. Furthermore, we determined the status of MSI, MLH1 promoter methylation, and other molecular alterations in separately microdissected intra-tumoral components and matched metastatic lymph nodes with heterogenous MMR expression by PCR and targeted NGS. As a result, heterogeneity of histomorphology in intratumoral components and metastatic lymph nodes was determined in 75% (3/4) cases (Case 1, 2, 4). In addition, heterogeneity of MLH1 promoter methylation and/or other common driving gene mutations of CRC, such as KRAS, PIK3CA and so on, was identi ed in all the 4 CRCs. These results indicated that heterogeneity of MMR protein expression and MSI status in CRC is rare, but exists indeed. These results also suggested that it might be necessary to detect biomarkers in different intra-tumoral components and corresponding metastases with heterogenous MMR expression for subsequent treatment.
Histopathologically, heterogeneity of morphology in intra-tumoral components and metastatic lymph nodes was determined in 75% (3/4) of cases of CRC, which was exactly matched with the heterogeneous zones of MMR protein expression. Although the association between the heterogeneity of MMR protein expression and heterogenous histomorphology was formally proposed in our study, it was also found in previous publications [13,16].  [24]. These results indicated that histomorphological heterogeneity might be an important clue for the heterogeneity of MMR protein expression and MSI status, and detection of MSI status only in one histomorphological component for these tumors might lead to misdiagnosis [25].
It was worth noting that the consistency of MSI status between primary and metastatic lesions was still controversial [15,[26][27][28]. Some scholars believed that metastatic tissue can be used to screen for Lynch syndrome and dMMR, because all cases examined showed consistent IHC staining between the primary tumors and metastases [27]. However, other scholars hold that the MMR protein expression in the metastatic tissue is not entirely consistent with that in the primary tumor, and the incidence of MSI in metastatic tissue is low [15,29,30]. In our study, the metastatic lymph nodes in Case 2 and 3 all showed glandular and cribriform structures with pMMR. These ndings suggested that the MSI status of the primary tumor could not be completely represented by the metastatic tissue, and provided strong evidence that the subclones with pMMR/MSS might be more prone to develop metastasis, consistent with previous publications [15,31]. Therefore, it is essential to detect the MMR protein expression and MSI status of metastases before treatment with chemotherapy and/or immune checkpoint inhibitors[6, 7], especially for the CRCs with intra-tumoral heterogenous MSI status.
The NGS analysis not only offers further evidence of the true MSI heterogeneity in the 4 cases, but also provides insights into the tumor biology. Loss of MMR protein expression is mainly caused by several reasons: (1) germline mutations of MMR genes; (2) methylation of the MLH1 gene promoter; (3) double somatic mutations of MMR genes [32,33]. In Case 1, germline pathogenic missense mutation of MLH1 (p.A681T) [34] was determined in both tumor zones with dMMR or pMMR. The heterogenous MMR protein expression, in this case, might be explained through allele frequency (AF) of MLH1 germline mutation, which suggests homozygous in Zone B with dMMR (AF=80.25%) and heterozygous in Zone A with pMMR (AF=48.98) ( In addition, heterogeneity of mutations in oncogenes (KRAS and PIK3CA), tumor suppressor genes (APC, PTEN, and SMAD4), and homologous recombination repair genes (BRCA2 and ATM) was determined in Case 1 and Case 4 with heterogeneous MMR protein expression and heterogeneous histomorphology, but not in Case 3 with homogeneous histomorphology, which provided strong evidence that these gene mutations also exist heterogeneity in CRC, as previous publication [39][40][41][42]. These results also indicated that histomorphological heterogeneity provided an effective clue for distinguishing the CRCs with heterogenous MSI status and other driver gene mutations.
In conclusion, heterogeneity of MMR protein expression and MSI status in CRC is rare but exists indeed. Our results provide strong evidence that the true heterogeneous MSI status is highly correlated with histomorphological heterogeneity, which is also an important clue for the heterogeneity of drive gene mutations in CRC, including KRAS, APC, PTEN, and PIK3CA. These results indicate that it is essential to detect the MMR protein expression and other gene mutations in metastases before the treatment with immune checkpoint inhibitors and/or other target drugs, especially for the CRC with heterogenous MMR protein expression. These results also emphasize the importance of histopathological morphology in the age of molecular pathology. However, because of limited cases in this study, heterogeneity of MMR protein expression and MSI status in CRC remains to be further research.   Supplementary Files