Next-Generation-Sequencing of the Human B-Cell Receptor Improves Detection and Diagnosis and Enhances Disease Monitoring in Patients with Gastric Mucosa-Associated Lymphoid Tissue Lymphoma
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsA detailed examination of the clonotypes that appear (or don't appear) with the patients raises a few questions that might be clarified. 1). The detection frequency for the NGS method is 88% (4 undetected in 36 samples) which is a little lower than the 93% in Multiple myeloma as you are aware (Pastushok et al., 2024). Explanation for why the 4 are undetected? Could it be the low B versus T cell content in these particular samples? please comment on this. 2). The higher frequencies of IghV3-11 and IghV4-34 did not go unnoticed, particularly since these families were also found to be abnormally frequent in the Multiple myeloma study. Is this a reflection of the pathogenicity or a bothersome over-representation created by the assay. Please comment.
Figure 3- Pie chart is not very clear. Could use more distinctive patterns such as color or black, clear and hatched.
Author Response
Reviewer 1:
Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.
Comment 1: A detailed examination of the clonotypes that appear (or don't appear) with the patients raises a few questions that might be clarified. The detection frequency for the NGS method is 88% (4 undetected in 36 samples) which is a little lower than the 93% in Multiple myeloma as you are aware (Pastushok et al., 2024). Explanation for why the 4 are undetected? Could it be the low B versus T cell content in these particular samples? please comment on this.
Response 1: Thank you for pointing this out. I am hesitant to believe that this could be due to the low percentage of B-cells in the samples because we observe that these 4 samples have a B-cell content ranging from 20% – 90%. It could be that at the point of sampling, the patients did not possess the malignant clonotype in the examined tissue section within levels detectable by this method, or that the biopsied area utilized for tissue sectioning had no tumor infiltrates, correlating with the histology results of the patients.
Comment 2: The higher frequencies of IghV3-11 and IghV4-34 did not go unnoticed, particularly since these families were also found to be abnormally frequent in the Multiple myeloma study. Is this a reflection of the pathogenicity or a bothersome over-representation created by the assay. Please comment.
Response 2: It is true that in two of the patients examined, the IGH V-gene usage of the malignant clonotype was IGHV4-34 and in one patient was IGHV3-11. However, I would not conclude that this gives any information about pathogenicity. For this, we would need to examine the IGHV-gene usage of a larger cohort of patients to make a detailed analysis and conclusion. Furthermore, the IGHV-gene usage in these samples must be compared to the IGHV-gene usage in normal (non-tumor) samples before a conclusion about pathogenicity can be made. This comparison is however, is not within the scope of this work.
Comment 3: Figure 3- Pie chart is not very clear. Could use more distinctive patterns such as color or black, clear and hatched.
Response 3: True. We have modified the figure accordingly. The modified figure can be found in the manuscript file, page 12.
Reviewer 2 Report
Comments and Suggestions for AuthorsThe authors reported the value of NGS analysis in G-MALT diagnosis and monitoring compared to histological analysis. The manuscript evaluates a relevant issue but some concerns were raised during the review process that needs to be clarified by the authors.
1. In the cohort of patients included in the study will be important to refer, if available, to the genetic translocations associated with each patient and if that could influence the obtained results. Furthermore, will be important to refer to the augment of B-cell infiltration in each sample, and how this correlates with NGS results. The authors should identify and discuss the resolution of each technique.
2. The authors must include a statistical analysis of the correlation or not between both techniques under evaluation, and the value of IGH and IGL independently.
3. Moreover, since different tissues were analyzed, the authors should analyze if the tissue sample influences the detection capacity and also integrate the therapy response.
4. The discussion needs to be improved and complemented with therapy response and location. Additionally, the authors must discuss better the H. pylori infection rate in the cohort and mention if the only patient positive was or not responsive to treatment.
5. Furthermore and align with point 3. The authors must discuss the relevance of IGH and IGL evaluation or if only one of them will be as informative as both, and also better explain the impact of somatic hypermutations on presented results.
6. The authors must revise the discussion related to CLL (lines 243-250) and support the arguments presented with references.
7. Overall the figures and the tables need to be revised to be easier to read and interpret, improving aspects, colors, and layout. The figure caption to not need to be included in the figure itself and if so it should be in the lower part.
Comments on the Quality of English Language
Overall the manuscript needs to be revised to improve some terms used and correct some typos, as in Table 1 “ BM infilteration”.
Author Response
Reviewer 2
Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.
Comment 1. In the cohort of patients included in the study will be important to refer, if available, to the genetic translocations associated with each patient and if that could influence the obtained results. Furthermore, will be important to refer to the augment of B-cell infiltration in each sample, and how this correlates with NGS results. The authors should identify and discuss the resolution of each technique.
Response 1: Thank you for pointing these observations out. The genetic translocations in these patients are unavailable to us at this time. The B-cell infiltration of the tumor samples are included in the supplementary file (see supplementary table 1, column D). With a higher percentage of B-cells, we expect to detect a broader B-cell repertoire in general and the tumor specific B-cell clone, if present, would be detected in the sample.
The samples we analyzed for each patient are samples with an unequivocal diagnosis of G-MALT at least at one time-point. NGS has a much better resolution and can be used to track the clonotypes due to its advantage of detecting nucleotide sequences of the tumor-specific clonotype. This is a feature that histology examination fails to provide. We have hence modified the text of the sentence in the discussion section on page 14, line 249 – 251: “NGS provides a robust method for the detection of clonal and sub-clonal populations of malignant cells and for disease monitoring owing to its a low limit of detection, high resolution and high specificity in comparison with other non-NGS based methods including histology.”
Comment 2. The authors must include a statistical analysis of the correlation or not between both techniques under evaluation, and the value of IGH and IGL independently.
Response 2: This is a good point. Statistical data has been included in the manuscript in section 3.1, page 10 lines 193 – 196: “Also, in 15/19 (78%) samples where histology examination suggested no evidence of lymphoma, we detected the tumor-specific clonotype by means of NGS (Figure 3). Moreover, in 4/19 (21%) samples without evidence of G-MALT based on histology data, we were also unable to detect the malignant clonotype using NGS (Figure 3).”
Concerning the value of IGH and IGL independently, we have done the analysis and included it in the manuscript page 14 to 15, line 271 to 277: “Moreover, based on IGH analysis, clonality was detected in 26/32 (81%) samples. IGL analysis revealed clonality in all 32 samples (100%). Our ability to detect clonality in IGL even when it was undetectable in IGH using NGS underscores the necessity of incorporating IGL assessments alongside IGH analysis during clonality testing. It's crucial to note that relying solely on IGH analysis may lead to false negative results where the clonal population may be missed during analysis. This is because the more frequent somatic hypermutations (SHMs) in IGH genes can impede efficient primer binding to the target sequences (28).”
Comment 3. Moreover, since different tissues were analyzed, the authors should analyze if the tissue sample influences the detection capacity and also integrate the therapy response.
Response 3: Good comment. Sampling procedures were performed according to standardized guidelines. In all cases, macroscopically most conspicuous lesions were sampled for initial diagnosis and follow-up. All biopsies were performed by a team of specialized gastroenterologists and radiologists. Therefore, the risk of a significant variance or sampling error based on the biopsy procedure appears to be very low. The type of tissue sample itself does not alter the detection capacity for clonal rearrangements. Treatment response information was not included in our study.
Comment 4. The discussion needs to be improved and complemented with therapy response and location. Additionally, the authors must discuss better the H. pylori infection rate in the cohort and mention if the only patient positive was or not responsive to treatment.
Response 4. Thank you for pointing this out. Yes, the patient was responsive to treatment but relapsed in 2010. We have thus included in table 1, patient D under the heading therapy for G-MALT, “effective antibiotic therapy against Helicobacter pylori”. We have also included for patient D in the table 1 under the heading comment on relapse “Relapse 08.09.2010”. Furthermore, we have edited our discussion on the patient on page 16, line 319 to 325: “There have been many reports on the strong association and frequency of occurrence of H. pylori infection in patients diagnosed with G-MALT (37-40). The incidence of Helicobacter pylori-negative MALT-Lymphoma from chronic gastritis was historically approximately 20% but has been increasing consistently over the past 20 years (41). Out of the 6 patients examined in our study, only one was positive with H. Pylori. The infection being initially absent from the patient following initial diagnosis of G-MALT. Therapy against H. pylori was associated with clinical remission in this patient in agreement with previous reports (5, 42).”
Comment 5. Furthermore and align with point 3. The authors must discuss the relevance of IGH and IGL evaluation or if only one of them will be as informative as both, and also better explain the impact of somatic hypermutations on presented results.
Response 5: In alignment with point 2, we have discussed the relevance of IGL in complementing IGH analysis. This can be found in the discussion section on page 14, lines 272 – 277: “Our ability to detect clonality in IGL even when it was undetectable in IGH using NGS underscores the necessity of incorporating IGL assessments alongside IGH analysis during clonality testing. It's crucial to note that relying solely on IGH analysis may lead to false negative results where the clonal population may be missed during analysis. This is because the more frequent somatic hypermutations (SHMs) in IGH genes can impede efficient primer binding to the target sequences”.
Comment 6. The authors must revise the discussion related to CLL (lines 243-250) and support the arguments presented with references.
Response 6: Thank you for this comment. We have revised the discussion related to CLL and the changes can be found on page 15, lines 304 – 312: “This could mean that the malignant clone in the gastrointestinal tissues are expanding via blood. Besides, although the co-occurrence of more than one clonally distinct B-cell lymphoid malignancy is a very rare but possible phenomenon, in one case, G-MALT was shown to be a secondary event in a CLL patient (33). In another patient, CLL was diagnosed 16 months following the initial diagnosis of G-MALT (34). Whether the clone which we detected was that which resulted in the CLL diagnosis remains to be clarified. However, the high frequency of this clonotype in the blood (21% - 89%) in comparison to the gastrointestinal tissues (0.4% - 9.2%) could be reflective of the higher number of tumor B-cells in blood relative to the tissues.”
Comment 7. Overall the figures and the tables need to be revised to be easier to read and interpret, improving aspects, colors, and layout. The figure caption to not need to be included in the figure itself and if so it should be in the lower part.
Response 7: Thanks for your suggestion. We have taken note of this and have adjusted the figures accordingly.
Reviewer 3 Report
Comments and Suggestions for Authors
The manuscript aims to enhance diagnostic capabilities by profiling BCR clonality in gastric mucosa-associated lymphoid tissue (G-MALT) lymphomas and correlating the frequency of clones detected by the BCR Pan-Clonality Assay with clinical sample diagnosis. Initially, the authors observed that all samples diagnosed as G-MALT based on histological or clinical analysis exhibited IGH/IGL rearrangement. Furthermore, even in samples lacking histological changes, clonal arrangements were detectable. Additionally, this method facilitated the detection of somatic hypermutation, providing valuable insights for disease interpretation.
Overall, the manuscript presents an intriguing diagnostic tool for precision medicine.
Comments:
-
The manuscript employs 30 ng of input gDNA for library preparation, contrary to the manufacturer's recommendation of 200 ng–2.0 µg according to Oncomine BCR Pan-Clonality Assay instructions. Could the authors elucidate the rationale behind this choice?
-
In the methods section, the authors referred to HE and IHC staining, yet the paper lacks the display of HE and IHC staining results. If the authors did not intend to demonstrate staining, the section regarding tissue staining and preparation should be omitted. Instead, authors can elaborate on gDNA extraction from formalin-fixed, paraffin-embedded tissue sections.
-
Have the authors observed any biased IGHV gene usage in the detected samples?
-
Supplementary Table 1 reveals that some samples exhibit very low reads (e.g., Patient C, 2007/01). Could the authors provide an explanation for this observation?
-
Is there a potential correlation between the frequencies of clones and disease progression?
Author Response
Reviewer 3
Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.
Comment 1: The manuscript employs 30 ng of input gDNA for library preparation, contrary to the manufacturer's recommendation of 200 ng–2.0 µg according to Oncomine BCR Pan-Clonality Assay instructions. Could the authors elucidate the rationale behind this choice?
Response 1: Thank you for this comment: Indeed the protocol recommended about 200ng input DNA. In many of our clinical samples, the DNA input amounted to 30ng. We discussed this with the manufacturers of the assay, who then supplied us with a supplementary protocol for samples in which the DNA input quantity is below 200ng. Information about this supplementary protocol is included in the manuscript page 8, line 167: “*Additional manufacturer's protocol for samples with low input quantity”. Furthermore, this information is included in the text within section 2.3 Library preparation, sequencing and clonality analysis, page 7, lines 153 – 154: “Library enrichment following the standard protocol for library amplification and 2x purification was also performed”. The protocol itself can be found on the manufacturer’s website 'Oncomine Human Immune Repertoire User Guide BCR Pan-Clonality Assay' on page 49.
Comment 2: In the methods section, the authors referred to HE and IHC staining, yet the paper lacks the display of HE and IHC staining results. If the authors did not intend to demonstrate staining, the section regarding tissue staining and preparation should be omitted. Instead, authors can elaborate on gDNA extraction from formalin-fixed, paraffin-embedded tissue sections.
Response 2: Thanks for this comment. HE and IHC staining were done for clinical diagnostics. Our results are hence based on information from the clinical reports of these patients with additional review for this project by experienced pathologists. We thus, do not intend to demonstrate the staining results. However, we would like to retain the section on tissue staining and preparation because, it is important to highlight how we determined the percentage of B- and T-cell content of the tissue samples. gDNA extraction was done using Maxwell, following the manufacturer's protocol which can be found on the manufacturer's website. To this effect, I have added in the manuscript, section 2.2 Tissue sectioning, staining and gDNA extraction, page 7 lines 147 - 148 “and following the manufacturer's instructions” for more clarity.
Comment 3: Have the authors observed any biased IGHV gene usage in the detected samples?
Response 3: The Authors have not identified any skewed IGHV gene usage in the given examples as the detected malignant clonotypes possessed diverse IGHV gene segments.
Comment 4: Supplementary Table 1 reveals that some samples exhibit very low reads (e.g., Patient C, 2007/01). Could the authors provide an explanation for this observation?
Response 4: Thanks for pointing this out. The low number of reads observed could be due to the low number of detectable B-cells in the sample. As shown in the manuscript table, the patient was administered with 6 cycles of Rituximab which targets B-cells leading to depletion. Another reason may be due to disparity in the quantity of the sample during the pooling step. However, we have seen that even with this low number of reads, the malignant clonotype was detected in the samples which further highlights the sensitivity of this method.
Comment 5: Is there a potential correlation between the frequencies of clones and disease progression?
Response 5: Good question but this is not the case.
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsThe author's reply did not entirely answer the raised concerns and further improvements need to be performed, namely:
1. The introduction of percentages in the results section did not mean that statistical analysis was performed. When comparing two techniques, a comparison analysis and concordance analysis must be performed to compare both techniques correctly. Furthermore, if statistical analysis was performed, the authors must include a section in the material and methods describing the test used and analysis details.
2. The quality of the sample collection was not called into question in the review process. The reviewer considered that the tissue of origin and its impact, if present, on the detection capacity should be explored, statistically analyzed, and included in the article for greater robustness of the work.
3. As previously mentioned, the layout of tables and figures must be improved to make them easier to read and interpret, improving aspects, colors, and layout.
Author Response
Reviewer 2
Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.
Comment 1. The introduction of percentages in the results section did not mean that statistical analysis was performed. When comparing two techniques, a comparison analysis and concordance analysis must be performed to compare both techniques correctly. Furthermore, if statistical analysis was performed, the authors must include a section in the material and methods describing the test used and analysis details.
Response 1: Thanks for this point. Statistical analysis has been done comparing both histology and NGS as well as IGH and IGL. We have included in the materials and methods section of page 5 to now read:
“2.4 Statistical analysis:
McNemar’s test for paired data was used for statistical analysis to compare the difference in sensitivity between histology versus NGS as well as to examine the correlation between clonality detection by analyzing IGH versus analyzing IGL. P-values < 0.05 were considered statistically significant.”
Furthermore we have included in the results section 3.1 on page 7, in the first paragraph “The difference between the sensitivity of the NGS method applied and histology was statistically significant (p = 0.0003).” The discussion section on page 11 was also edited to read “The difference in the number of samples in which clonality was detected in IGL as compared to IGH was statistically significant (p = 0.0005).”
Comment 2. The quality of the sample collection was not called into question in the review process. The reviewer considered that the tissue of origin and its impact, if present, on the detection capacity should be explored, statistically analyzed, and included in the article for greater robustness of the work.
Response 2: Thanks for this comment. The impact of the tissue of origin on the detection capacity of the assay has been explored and there is no correlation. We would like to add that the sampling of tissues from different body sites was not to compare the detection capacity of the technique but to check for early signs of tumor metastases or extensive body site involvement which was missed by histology. This is what is meant in the second to the last paragraph of the discussion section on page 12: “That we detected the malignant clonotype in nearby tissues by means of NGS where this was missed out based on histology shows the usefulness and accuracy of this technique in early detection of malignancy, and nearby tissue involvement, which could assist in timely institution of further relevant therapeutic strategies to inhibit spread of the tumor.”
Comment 3. As previously mentioned, the layout of tables and figures must be improved to make them easier to read and interpret, improving aspects, colors, and layout.
Response 3: Thanks. They have been improved.
Round 3
Reviewer 2 Report
Comments and Suggestions for AuthorsThe authors reply to the reviwers issues.