Loss of CREBBP and KMT2D cooperate to accelerate lymphomagenesis and shape the lymphoma immune microenvironment

Despite regulating overlapping gene enhancers and pathways, CREBBP and KMT2D mutations recurrently co-occur in germinal center (GC) B cell-derived lymphomas, suggesting potential oncogenic cooperation. Herein, we report that combined haploinsufficiency of Crebbp and Kmt2d induces a more severe mouse lymphoma phenotype (vs either allele alone) and unexpectedly confers an immune evasive microenvironment manifesting as CD8+ T-cell exhaustion and reduced infiltration. This is linked to profound repression of immune synapse genes that mediate crosstalk with T-cells, resulting in aberrant GC B cell fate decisions. From the epigenetic perspective, we observe interaction and mutually dependent binding and function of CREBBP and KMT2D on chromatin. Their combined deficiency preferentially impairs activation of immune synapse-responsive super-enhancers, pointing to a particular dependency for both co-activators at these specialized regulatory elements. Together, our data provide an example where chromatin modifier mutations cooperatively shape and induce an immune-evasive microenvironment to facilitate lymphomagenesis.

The paper presented by Li and colleagues describes the phenotypic changes induced by combined GC-specific haploinsufficiency of the histone/chromatin modifier genes Crebbp and Kmt2d.The project has a high clinical relevance as the vast majority of lymphoma patients harbor inactivating mutations in both genes.
Based on their in vivo systems (using conditional mouse mutants) the authors gain insights in B cell lymphoma pathogenesis and provide potential rationalities for the development of enhancer targeting approaches to treat lymphoma patients.
Overall, the presented data is convincing and the manuscript is well written.However, the following limitations apply: Major concerns: 1) The laboratories of Riccardo Dalla-Favera and Laura Pasqualucci recently published their findings on combinatorial Crebbp and Kmt2d inactivation in GC B cells and human DLBCL (Vlasevska et al.PNAS 2023).The authors' discussion about the competing manuscript is missing.
2) The authors' lymphoma development studies (based on a Bcl2 transgenic background) are unique to the current study and their findings are very interesting.A detailed characterization of the aberrant B cell population will be even more informative: The authors define lymphoma development by the histological appearance of the expanded cells.In addition, BCR clonality studies (by VDJ sequencing approaches) in the expanded B cell pool will discriminate between polyclonal B cell expansion (typically evident in Bcl2 transgenic mice) and mono-/oligoclonal tumor development.In representative tumors Li and colleagues might want to confirm Cre-mediated combined loss of Crebbp and Kmt2d (e.g. by PCR based detection of exon loss in lymphoma cells).
The transcriptional and (epi)genetic profiling of the mono/oligoclonal tumor clones detected in the various genotypes (and the comparison of these mouse findings with human FL data sets) would be very interesting as "tertiary" hits and insights in lymphoma evolution could be identified.
Beside the spleen do the authors observe B cell expansion in other lymphoid sites (e.g.peripheral LN, mes LN, PP, BM)?In the absence of transgenic Bcl2 overexpression, do the (aged) animals suffer from tumor development induced by GC-specific haploinsufficiency of both epigenetic modifiers?
The survival curve implies a similar disease course in BCL2+K and BCL2+CK mice, whereas BCL2+K animals will reach a survival plateau.Are these animals lymphoma-free or do they suffer from less-aggressive lymphoma entities?
3) The lack of cytotoxic T cell infiltration represents a key feature in BCL2+CK tumors compared to other genotypes.In contrast, CD4+ cells that form the immunological synapse required for B cell selection during the GC reaction, seem to be unaffected, at least their quantity in the tumor samples is independent of the underlying genotype.A more detailed characterization of the T cell pools infiltrating the lymphoma area (e.g.activation status, exhaustion marker expression, clonal selection) would be helpful to better understand the mechanisms of immune evasion which will take place during lymphoma development and progression.Furthermore, the authors have the unique opportunity to study T -B cell interactions in their autochthonous lymphoma models and may add some functional assays.4) Crebbp and Kmt2d co-localize in a complex and impact on its expression levels and function by post-translational modifications (see Vlasevska et al.PNAS 2023).
The authors might want to prove the conservation of this interplay in their mouse tumors, e.g by determining Crebbp (and its paralog p300?) and Kmt2d protein expression and their modifications in the tumors.

Minor concerns:
Please report the percentage of combined mutations affecting Crebbp and Kmt2d in the data set used for Suppl.Figure 1

Reviewer #2 (Remarks to the Author):
Li and colleagues conducted in vivo experiments to investigate the mechanisms of the concurrent mutations in CREBBP and KMT2D in both GC development and lymphomagenesis.They observed that the loss of CREBBP and KMT2D leads to a more severe lymphoma phenotype and unexpected immune evasion behavior in cancer cells.This loss results in reduced infiltration of CD8+ T cells and suppression of immune synapse genes, contributing to a weakened T-cell response and flawed cell fate decisions, possibly aiding cancer cells in eluding immune surveillance.Additionally, from an epigenetic standpoint, the loss of cooperation between C+K had a significant and profound effect on super enhancers, especially those responsible for regulating immune synapse signaling genes.Functional characterization of the mutations identified in lymphoma derived from humans is essential for a more comprehensive understanding of these genes' roles in lymphomagenesis.The manuscript is well-crafted, and some comments are listed below: 1. On page 6, lines 3-5, the authors calculated the p-values for the co-occurrence of CREBBP and KMT2D mutations in FL.However, the text does not specify the number of genes that were selected for this calculation.Moreover, it remains unclear whether the authors took into account the potential functional effects of the mutations, such as whether they were damaging or not, in their analysis.
2. In Supplementary Fig. 1d, the authors concluded that BCL2+CK exhibited 'highly, larger, more' features compared to others.However, the conclusion lacks statistical support, leaving the basis for this assertion unclear.
3. On page 7, line 11, the authors noted that the SHM (somatic hypermutation) burden varied among different genotypes.However, they did not specify which isotype of VH was analyzed.Understanding whether the VH isotypes were class-switched (e.g., IgG) or not (e.g., IgM) would be relevant, as it could significantly influence the SHM levels.
4. In the case of human FL patients, the authors do not make it clear whether they observed any differences in prognosis among the C, K, CK, and WT patients.Likewise, the manuscript lacks information on whether similar conclusions were reached regarding CD8+ T cells in the different mutation groups of FL or DLBCL.
5. The authors concluded that the loss of function in CREBBP and KMT2D cooperates to accelerate FL development, manifesting more aggressive characteristics than either allele alone.However, Figure 1b and 1d do not visibly differentiate between BCL2+k and BCL2+CK.
An explanation from the authors may be needed to clarify this apparent inconsistency.6.It is unclear from the manuscript whether the authors examined the phenotype of CD8 or CD4 T cells in the BCL2+CK mice or human FLs, such as signs of exhaustion or hyperactivity 7. On page 7, the flow data reveals a similar abundance of CD4+ T cells but a reduction in CD8+ T cells in BCL2+CK mice.It would be crucial for the authors to specify which CD8+ cell subtypes were reduced, as this information could enhance the understanding of the immune response dynamics.8. On page 16, line 2, the authors performed RNA sequencing on B220+ cells, identifying them as lymphoma cells.However, the manuscript does not specify the criteria used to define these cells as tumor cells.9. On page 18, the authors mentioned impaired expression of key genes involved in GC B cell interaction with TFH cells and GC exit.However, how C+K differentially modulate the infiltration of CD8+ and CD4+ T cells remains unclear, given that BCL2+CK mice mainly had reduced CD8+ T cells rather than CD4+ T cells.Further clarification on this aspect is needed.

Reviewer #3 (Remarks to the Author):
In this interesting manuscript, Li and colleagues investigate the effects of combined heterozygous loss of CREBBP and KMT2D in lymphoma, using the VavP-Bcl2 model and the GC B-cell specific Cg1-Cre strain.Using a variety of in vitro and in vivo analyses, authors show that combined loss of CREBBP and KMT2D induces a more severe phenotype than loss of each of them separately.This is associated to a phenotype of immune evasion with drastically reduced infiltrating CD8 T-cells.Furthermore, authors show the cooperative nature of CREBBP and KMT2D in the regulation of their epigenetic target programs, and even demonstrate that CREBBP and KMT2D co-interact.Finally, the epigenetic phenotype upon combined KMT2D/CREBBP loss was especially strong at superenhancer regions (and, more specifically, at super enhancers driving the expression of immune synapse signaling genes).Overall, this is a very comprehensive study uncovering novel relevant biology, and I would like to congratulate authors for their really nice work.
My only minor comment to authors is that it would be good if they could show all 4 genotypes in Fig 4i (as of know, it is not clear if the anti-CD40 bar is grouping all 4 genotypes, or if it's just one genotype only).Similarly, it would be good if authors could show the other 2 genoypes in Fig 4j, beyond the WT and CK, so that we can evaluate potential differences between single and combined genotypes.

REVIEWER COMMENTS
Reviewer #1 (Remarks to the Author): The paper presented by Li and colleagues describes the phenotypic changes induced by combined GC-specific haploinsufficiency of the histone/chromatin modifier genes Crebbp and Kmt2d.The project has a high clinical relevance as the vast majority of lymphoma patients harbor inactivating mutations in both genes.Based on their in vivo systems (using conditional mouse mutants) the authors gain insights in B cell lymphoma pathogenesis and provide potential rationalities for the development of enhancer targeting approaches to treat lymphoma patients.Overall, the presented data is convincing and the manuscript is well written.However, the following limitations apply: Answer: We thank the reviewer # 1 for these positive comments on our work.

Answer:
We thank the reviewer for this comment.We have added the following text to the discussion (all additions to the manuscript are in blue text): "a recent study also observed germinal center hyperplasia with CK double loss of function, and demonstrated that CREBBP forms a complex with KMT2D and directly acetylates KMT2D to modulate KMT2D activity at gene enhancers 1 , further emphasizing the various facets of cooperation between CREBBP and KMT2D in GC B-cells."R1.2.The authors' lymphoma development studies (based on a Bcl2 transgenic background) are unique to the current study and their findings are very interesting.A detailed characterization of the aberrant B cell population will be even more informative: The authors define lymphoma development by the histological appearance of the expanded cells.In addition, BCR clonality studies (by VDJ sequencing approaches) in the expanded B cell pool will discriminate between polyclonal B cell expansion (typically evident in Bcl2 transgenic mice) and mono-/oligoclonal tumor development.In representative tumors Li and colleagues might want to confirm Cre-mediated combined loss of Crebbp and Kmt2d (e.g. by PCR based detection of exon loss in lymphoma cells).

Answer:
We thank the reviewer for these suggestions.
1) Regarding the first of these two points, we performed Ig heavy chain VDJ region sequencing to better assess lymphoma clonal architecture.Overall, BCL2+CK lymphomas tend to comprise fewer BCR-distinct clones than BCL2 (Revision Fig. 1a).A similar trend of clone number reduction was observed in BCL2+C and BCL2+K, although to a lesser degree.Consistently, a more quantitative measure of BCR clonality by calculating either Simpson clonality score or Shannon diversity score revealed that BCL2+CK lymphomas tend to be more clonal and accordingly less diverse than BCL2, with BCL2+C and BCL2+K adopting an intermediate phenotype (Revision Fig. 1b-c).We included this new data in the revised manuscript as Supplementary Fig. 2a-b 2) Regarding the second part of the reviewer's question, we have confirmed the Cre-mediated heterozygous exon deletion of Crebbp and/or Kmt2d in day 235 murine lymphoma samples by genotyping PCR (Revision Fig. 2), which is also included as the new Supplementary Fig. 1b-c.

Answer:
We thank the reviewer for raising this interesting point.We have some experience performing exome sequencing on various lymphoma models and have not found these to be informative, since there are never enough numbers of distinct individual mice to show statistically significant patterns (these patterns are only evident when hundreds of human patients are profiled).Moreover, tertiary hits in mice have particularities that may be related to how lymphomas progress in the context of model systems with many competing clones.Regardless, in appreciation of the reviewer's interest in this topic we performed whole exome sequencing of day 235 murine lymphoma samples (4 replicates per genotype).After filtering out putative germline variants, defined as SNVs (single nucleotide variants) shared among two or more replicates, we obtained 24 highly enriched gene mutations (VAF>0.3)among all samples (Revision Table 1).As we anticipated, the biological impact of these mutations remains unclear and as such we did not include these data in the revised manuscript.Answer: We thank the reviewer for this question.To address this point we performed flow cytometry analysis of frozen bone marrow cells from our lymphoma mice cohort.As expected, the BCL2 allele resulted in increased abundance of B220+ B-cells across genotypes (Revision Fig. 3a-b), given the role of BCL2 in expanding B cell populations.Notably this effect was most evident at day 116 but not day 235 post BMT, for reasons that are currently unclear.Notably, the B220+ cell expansion in BCL2+CK was significantly higher than that in other genotypes (Revision Fig. 3a-b).To further understand these findings we next checked whether the observed B220+ cell expansion was derived from invading GCB-like cells, which more directly reflects the lymphoma phenotype.Using markers of normal GC B-cells (B220+CD38-FAS+) we observed enrichment for this population among all BCL2 containing genotypes as compared to WT (Revision Fig. 3a,  3c).However, lymphomas do not necessarily express identical markers as normal GC B-cells, and along these lines we observed a trend towards enrichment of GCB-like cells (B220+/FAS+) among the CK lymphomas vs the other BCL2 containing genotypes (Revision Fig. 3a, 3d).This result suggests that lymphoma B cells likely migrate to and expand in BM.However these findings require further study to better understand the biology of these cells.Apart from bone marrow, we unfortunately didn't collect LNs from this cohort so cannot show any data on those tissues.Answer: We thank the reviewer for bringing up this interesting point.We didn't include CK haploinsufficiency alone mice (without BCL2 overexpression) in our murine lymphoma cohort.
Given that CK mutant human lymphomas virtually always carry BCL2 translocations, we reasoned that CK alone setting would not be as informative for studies of the human disease.Although it might be interesting conceptually, we decided to restrict our use of mice to those experiments that were most directly relevant to the physiological scenario, in part given concerns regarding excessive use of vertebrate organisms.
R1.6.The survival curve implies a similar disease course in BCL2+K and BCL2+CK mice, whereas BCL2+K animals will reach a survival plateau.Are these animals lymphoma-free or do they suffer from less-aggressive lymphoma entities?
Answer: We thank the reviewer for this interesting question.We didn't examine in further detail those few BCL2+K mice still alive at the end of our 300-day monitoring period.However, given that likely malignant GCB cell expansion was uniformly observed in randomly selected, healthylooking BCL2+K mice at day 235 post BMT (Fig. 1d), we speculate that those few late-surviving BCL2+K mice may have had lower grade lymphomas rather than being lymphoma-free.

R1.7.
The lack of cytotoxic T cell infiltration represents a key feature in BCL2+CK tumors compared to other genotypes.In contrast, CD4+ cells that form the immunological synapse required for B cell selection during the GC reaction, seem to be unaffected, at least their quantity in the tumor samples is independent of the underlying genotype.A more detailed characterization of the T cell pools infiltrating the lymphoma area (e.g.activation status, exhaustion marker expression, clonal selection) would be helpful to better understand the mechanisms of immune evasion which will take place during lymphoma development and progression.Furthermore, the authors have the unique opportunity to study T -B cell interactions in their autochthonous lymphoma models and may add some functional assays.

Answer:
We thank the reviewer for this excellent suggestion.To address the first point we performed Cytek multi-dimensional flow cytometry analysis of T-cell phenotypes (naive, central memory, effector, exhaustion, regulatory T cells, Tfh, Tfr) using the remaining frozen splenocyte specimens from our lymphoma cohort mice at both time points to gain more insights into how CK loss of function differentially affects CD4 and CD8 T cells in the lymphoma microenvironment.As noted before, CD8, but not CD4, displayed progressive reduction from BCL2+C, BCL2+K, to BCL2+CK, relative to BCL2 at both time points (Revision Fig. 4a-b).Given that all subsequent analyses were based on these specimens, we replaced the former flow cytometry analysis from Fig. 1g-h with these data, to directly match the more detailed flow cytometry results.On the CD8 side, at day 116 we observed significant expansion of CD8 effector cells (CD8+CD44+CCR7-), accompanied by a corresponding reduction of central memory (CM, CD8+CD44+CCR7+) CD8 cells in BCL2+CK, that was greater than effects observed in BCL2+C and BCL2+K (Revision Fig. 5a-b).
We note for the reviewer that day 235 specimens were actually thawed and examined in our lab prior to the day 116 tumors mentioned just above.Unfortunately, we were unable to accurately separate naive, CM and effector cells at day 235 due to failure of the CD62L channel, caused by the known pitfall of shedding and loss of this marker (that we were initially unaware of) after freeze-thaw cycle as previously reported 2 .This was why we switched to subsequent labeling of CCR7 in our day 116 samples.
This issue notwithstanding, at day 235 we noted significant expansion of CD44+ activated CD8 T cells (which contain both effector and CM populations) in BCL2+C and BCL2+CK, whereas the fraction of these cells in BCL2+K was reduced compared to BCL2 (Revision Fig. 5c-d).This interpretation is limited by the lack of CD62L staining at day 235 and so should not be viewed as conclusive.
Most remarkably, even though the abundance of CD8 cells was strongly reduced at day 235 in the BCL2+CK setting, there was a massive and significant increase in the proportion of these cells manifesting an exhausted phenotype, whereas the proportion of exhausted cells was roughly similar to BCL2 alone in the BCL2+C and BCL2+K setting (Revision Fig. 5e-f).In contrast there was no change in the abundance of exhausted CD8 cells at day 116.Together, these data suggest that CREBBP and KMT2D loss of function cooperatively remodels the immune microenvironment from a TFH-enriched pro-GC reaction early stage to a CD8exhausted and depleted later stage to facilitate lymphomagenesis (Revision Fig. 7).These new CD4 and CD8 flow data are included in the revised manuscript as Fig. 1g-j, 1o, Supplementary Fig. 2c-o.Regarding the reviewer's second point, given that we don't have ongoing lymphoma cohorts for functional studies of B-T interaction, as an alternative we performed in vitro co-culture assays by mixing irradiated isogenic OCI-Ly7 cells (WT and CK) pre-pulsed with either DMSO vehicle or CEF peptide (a pool of HLA class I-restricted virus peptides), with HLA-A matched human peripheral blood CD8 T cells at 1:1 ratio in the presence of IL2 and IL15 for 10 days, followed by FACS-mediated analysis of CD8 T cell activation and cytokine production (Revision Fig. 8a).We hypothesized that CK deficiency-induced reduction in immune synapse co-stimulatory proteins would result in weaker activation of CD8 cells.

Although we do not know the etiology of this phenotypic discovery, it does indicate significant CD8 T-cell dysfunction in the BCL2+CK setting. Lacking more specific information on the fraction of effector vs CM cells in this context we suggest it is best to avoid concluding at this point that the cause of exhaustion is linked to the apparent increase in activated CD8 T-cells. However, it does further underline how cooperative effects of C+K can alter the immune landscape and we thank the reviewer for encouraging us to add these interesting experiments to the manuscript.
As expected, CEF-pulsed OCI-Ly7 cells induced CD8 expansion to a greater degree than their DMSO-treated counterparts which presumably activated CD8 mainly through MHC-mismatchtriggered alloreactive response (Revision Fig. 8b-d).In both conditions, CK-deficient OCI-Ly7 cells were defective in activating and expanding CD8 T cells.However, it is noteworthy that this effect was more significant in the CEF treated setting, indicating a defect of CK cells in activating antigen-specific CD8 T-cell responses.
We next wondered whether this weaker stimulation results in defective T cell differentiation.Indeed, when zooming in to different CD8 subtypes, we observed an increase of naïve (hCD45RA+hCD62L+hCD95-) and central memory (CM, hCD45RA-hCD62L+), and a concordant decrease of effector memory (EM, hCD45RA-hCD62L-) or effector (hCD45RA+hCD62L-) cells upon CK stimulation than WT (Revision Fig. 8e-f).Finally, re-stimulation of CK-exposed CD8 cells yielded reduced induction of IFNg and TNFa (Revision Fig. 8g-i), two signature effector cytokines of cytotoxic CD8 cells involved in tumor clearance 5 , indicating that CK-stimulated CD8 cells were functionally less active.
These conditions allow dissection of isogenic WT vs CK T-cell stimulatory effects to be visualized at the single cell level with controlled ratios of cells and demonstrate impaired CD8 activation in vitro.This system enables observations to be made that cannot be dissected out from the syngeneic tumor microenvironment in vivo, thus providing a more complete picture of the immune perturbation induced by CK mutations, as requested by the reviewer.These new data are included in the revised manuscript as Fig. 6o-q and Supplementary Fig. 6p-v.The implications of our collective findings and further integration of these results are included in the revised discussion section.

Answer:
We thank the reviewer for this question.Significance of co-occurrence was evaluated using Fisher's exact test and only CREBBP and KMT2D mutations were evaluated.FL_all merged dataset in Supplementary Fig. 1a comprised the following cohorts: In total, FL_all contains 478 patients, out of which 250 (52.3%) carry CREBBP mutation, 305 (63.8%) carry KMT2D mutation, 185 (38.7%) carry both mutations.We have included these patient numbers and percentages in the revised Supplementary Fig. 1a legend.
Given that SIFT or PP2_HDIV_PRED based toxicity prediction is not available for these cohorts, we made no assumptions regarding the deleterious nature of CREBBP and KMT2D mutations when performing co-occurring analysis.

Answer:
We thank the reviewer for raising this concern.The reported histological analyses were performed by our co-author and internationally recognized hematopathologist Dr. Amy Chadburn, who is a member of the WHO lymphoma classification team 7 .Dr. Chadburn made this statement based on her observation that most BCL2+CK spleens exhibited these features.This is how such diagnostic determinations are made in routine clinical practice.Given that it is technically challenging to quantitatively measure cells exhibiting highly heterogeneous morphology, we have included this caveat statement: "based on detailed hematopathology reporting" in our revised text.
R2.3.On page 7, line 11, the authors noted that the SHM (somatic hypermutation) burden varied among different genotypes.However, they did not specify which isotype of VH was analyzed.
Understanding whether the VH isotypes were class-switched (e.g., IgG) or not (e.g., IgM) would be relevant, as it could significantly influence the SHM levels.
Answer: We thank the reviewer for pointing this out.Our PCR primers target Ig heavy chain VDJ region, which cannot distinguish different Ig heavy chain isotypes.To address this Ig isotype composition question, we checked the expression levels of different Ig constant genes (encoding different isotypes) from our day 235 murine lymphoma RNA-seq dataset.We found that the Ig isotype composition across all samples was rather heterogeneous, without bias towards any particular isotype selection (Revision Fig. 10), suggesting the observed heavier SHM burden is not likely due to changes in Ig class switching.In addition, the link between Ig class switching and SHM is elusive since Ig class switching recombination was shown to frequently occur prior to GC reaction 8 .We have included these data in the revised manuscript as Supplementary Fig. 1j.Answer: We appreciate the reviewer's suggestion.Unfortunately, there are still very limited instances of clinically annotated and sequenced FL patients, which makes it hard to address such issues in a statistically robust manner.Moreover, FL treatment is quite heterogeneous, may involve many different and sequential therapeutic regimens, and outcomes are further dependent on how tumors with particular mutational constellations specifically respond to any one of the various therapeutic options.These considerations significantly limit biological conclusions that can be drawn based on currently available data.
However, to gain at least some insight into this question we initiated a new collaboration for this revision with Dr. Oliver Weigert (Ludwig Maximilian University of Munich), who is a leader in the FL genetic biomarkers field.His group conducted an outcomes analysis using their published GLSG2000 trial dataset together will an additional British Columbia Cancer Agency (BCCA) FL dataset 9 .Importantly, most patients in these datasets had advanced FL requiring R-CHOP-like systemic treatment.As such, outcomes are based on more DLBCL-like FL stages which may or may not reflect outcomes as examined from more indolent stages of disease.
Because of these considerations we used failure free survival as an endpoint (very similar to PFS and TTF), since this endpoint is probably clinically most meaningful in this context.OS is heavily biased by subsequent therapies (especially in indolent diseases like FL).Also, given the very high response rates to R-CHOP-like regimens, ORR etc. is very unlikely to show differences and is clinically not relevant.This analysis revealed that patients carrying C, K, or CK mutations showed equivalent trends to their WT counterparts (Revision Fig. 11).
There are several caveats in this analysis that could explain the minor differences observed.1) The patient numbers are highly unbalanced among different groups due to the high frequency of C/K mutations in FL. 2) Disease in WT patients must be driven by other mutations, with distinct biological dependencies and functions, which will affect response to specific treatments.3) There is insufficient statistical power to consider multiple covariates such as mutation burden, gender, age, health history etc. among the different groups.
Finally, CD8 staining was not available for these cohorts and so could not be factored in, again even where they available such analyses would be affected by the same limitations mentioned above.Given all of this we have not included these findings in the revised manuscript, although Dr. Weigert's group is acknowledged for the time and effort they devoted to addressing this question.

Answer:
We understand the reviewer's concern.We agree that the difference between BCL2+K and BCL2+CK in Fig. 1b and 1d is minor, yet the trend that BCL2+CK is more deleterious than BCL2+K is clearly visible.Furthermore, we provide many other data (including new results from this revision) showing the more aggressive phenotype in BCL2+CK, such as histology, SHM, survival, reduced clonality, and CD8 reduction and exhaustion (Fig. 1e, 1f, 1h-n, Supplementary Fig. 2a-b).
R2.6.It is unclear from the manuscript whether the authors examined the phenotype of CD8 or CD4 T cells in the BCL2+CK mice or human FLs, such as signs of exhaustion or hyperactivity

Answer:
We thank the reviewer for this great suggestion.As mentioned for reviewer 1 (R1.7): We performed Cytek multi-dimensional flow cytometry analysis of T-cell phenotypes (naive, central memory, effector, exhaustion, regulatory T cells, Tfh, Tfr) using the remaining frozen splenocyte specimens from our lymphoma cohort mice at both time points to gain more insights into how CK loss of function differentially affects CD4 and CD8 T cells in the lymphoma microenvironment.As noted before, CD8, but not CD4, displayed progressive reduction from BCL2+C, BCL2+K, to BCL2+CK, relative to BCL2 at both time points (Revision Fig. 4a-b).Given that all subsequent analyses were based on these specimens, we replaced the former flow cytometry analysis from Fig. 1g-h with these data, to directly match the more detailed flow cytometry results.On the CD8 side, at day 116 we observed significant expansion of CD8 effector cells (CD8+CD44+CCR7-), accompanied by a corresponding reduction of central memory (CM, CD8+CD44+CCR7+) CD8 cells in BCL2+CK, that was greater than effects observed in BCL2+C and BCL2+K (Revision Fig. 5a-b).
We note for the reviewer that day 235 specimens were actually thawed and examined in our lab prior to the day 116 tumors mentioned just above.Unfortunately, we were unable to accurately separate naive, CM and effector cells at day 235 due to failure of the CD62L channel, caused by the known pitfall of shedding and loss of this marker (that we were initially unaware of) after freeze-thaw cycle as previously reported 2 .This was why we switched to subsequent labeling of CCR7 in our day 116 samples.
This issue notwithstanding, at day 235 we noted significant expansion of CD44+ activated CD8 T cells (which contain both effector and CM populations) in BCL2+C and BCL2+CK, whereas the fraction of these cells in BCL2+K was reduced compared to BCL2 (Revision Fig. 5c-d).This interpretation is limited by the lack of CD62L staining at day 235 and so should not be viewed as conclusive.
Most remarkably, even though the abundance of CD8 cells was strongly reduced at day 235 in the BCL2+CK setting, there was a massive and significant increase in the proportion of these cells manifesting an exhausted phenotype, whereas the proportion of exhausted cells was roughly similar to BCL2 alone in the BCL2+C and BCL2+K setting (Revision Fig. 5e-f).In contrast there was no change in the abundance of exhausted CD8 cells at day 116.Although we do not know the etiology of this phenotypic discovery, it does indicate significant CD8 T-cell dysfunction in the BCL2+CK setting.Lacking more specific information on the fraction of effector vs CM cells in this context we suggest it is best to avoid concluding at this point that the cause of exhaustion is linked to the apparent increase in activated CD8 T-cells.However, it does further underline how cooperative effects of C+K can alter the immune landscape and we thank the reviewer for encouraging us to add these interesting experiments to the manuscript.On the CD4 side, TFH (T follicular helper, CD4+CXCR5+PD1+FOXP3-) cell frequency was progressively increased from BCL2+C, BCL2+K, to BCL2+CK, relative to BCL2 at both time points (Revision Fig. 6a-c).Intriguingly, TFH expansion occurred out of proportion with GC suppressive T follicular regulatory cells (TFRs, CD4+CXCR5+PD1+FOXP3+) at day 116 (Revision Fig. 6de).Given that TFRs are known to terminate the GC reaction in part by suppressing TFH functions 3,4 , we speculate that increased TFH/TFR ratios may favor initial expansion of malignant GC-like structures.Consistently, TFH/GCB ratios were significantly increased in BCL2+K and BCL2+CK compared to BCL2 at day 116 (Revision Fig. 6f).Finally, there was an equivalent increase in a.The data in Fig 6 were generated predominantly in a single cell line (OCI-Ly7).The authors might want to use their murine lymphoma samples to validate selected findings in another model system.Have the authors acquired data for chronic GCs as detectable in mesLN (or PP)?Is class switch recombination undisturbed upon Crebbp and Kmt2d haploinsufficiency?Impaired IgG1 switching might impact on the results in Suppl Fig 3 e-j.

1 .
The laboratories of Riccardo Dalla-Favera and Laura Pasqualucci recently published their findings on combinatorial Crebbp and Kmt2d inactivation in GC B cells and human DLBCL(Vlasevska et al.PNAS 2023).The authors' discussion about the competing manuscript is missing.

. Revision Fig. 1 .
BCR immuno-seq (targeting Ig heavy chain VDJ regions) of day 235 murine lymphoma samples.a, Bar graphs depicting the number and fraction of all distinct IgH clones in each genotype.Each bar represents 1 mouse, 4 mouse replicates per genotype.b-c, Pie charts showing the fraction distribution of different IgH clonality (b) or diversity (c) categories in each genotype.The 25 th and 75 th percentiles were used as cut-off for intermediate vs low and high vs intermediate respectively.

Revision Fig. 10 .R2. 4 .
Stacked bar plots showing the fraction of different Ig heavy chain isotype genes in day 235 murine lymphoma samples based on RNA-seq data.In the case of human FL patients, the authors do not make it clear whether they observed any differences in prognosis among the C, K, CK, and WT patients.Likewise, the manuscript lacks information on whether similar conclusions were reached regarding CD8+ T cells in the different mutation groups of FL or DLBCL.

Beside the spleen do the authors observe B cell expansion in other lymphoid sites (e.g. peripheral LN, mes LN, PP, BM)?
Revision Table1.A list of genes carrying unique, high-severity de novo mutations in day 235 murine lymphoma samples as revealed by whole exome sequencing.The values represent VAFs for the indicated mutant alleles (row names) in the indicated samples (column names).*denotes different mutations on the same gene.R1.4.