Codon bias imposes a targetable limitation on KRAS-driven therapeutic resistance

KRAS mutations drive resistance to targeted therapies, including EGFR inhibitors in colorectal cancer (CRC). Through genetic screens, we unexpectedly find that mutant HRAS, which is rarely found in CRC, is a stronger driver of resistance than mutant KRAS. This difference is ascribed to common codon bias in HRAS, which leads to much higher protein expression, and implies that the inherent poor expression of KRAS due to rare codons must be surmounted during drug resistance. In agreement, we demonstrate that primary resistance to cetuximab is dependent upon both KRAS mutational status and protein expression level, and acquired resistance is often associated with KRASQ61 mutations that function even when protein expression is low. Finally, cancer cells upregulate translation to facilitate KRASG12-driven acquired resistance, resulting in hypersensitivity to translational inhibitors. These findings demonstrate that codon bias plays a critical role in KRAS-driven resistance and provide a rationale for targeting translation to overcome resistance. Supplementary information The online version of this article (doi:10.1038/ncomms15617) contains supplementary material, which is available to authorized users.

Overall, the link between codon usage and RAS expression is established through well-designed experiments in multiple cell lines, suggesting a general role for codon usage in regulating RAS expression. The observation that strongly-activating KRASQ61R mutations can drive resistance, despite low expression, while weakly-activating KRASG12/13 mutations require increased expression levels is very interesting, and suggests an novel mechanism of RAS-dependent drug resistance. However, the link between "translational upregulation" and increased KRAS levels is correlative. The manuscript would be greatly strengthened if the authors would provide evidence showing that the increased KRAS levels are the result of augmented protein synthesis, and are sufficient to confer drug resistance. Additionally, it is not clear if the lethality of drugs targeting the translational apparatus is due specifically to the decrease in KRAS levels, or due to either an overall decrease in protein synthesis or other targets affected by the compounds employed in the experiments. In principle, the manuscript is suitable for publication in Nature Communications, but these points should be addressed prior to acceptance.
Major comments: 1. In Figure 5, the authors show a correlation between global protein synthesis rates and KRAS expression, but it is not clear if the relationship between the two is causal. The observation that mTOR inhibitors abrogate KRAS expression implicates increased translation initiation and/or elongation rates in driving KRAS expression. However, mTOR kinase phosphorylates many targets including components of the protein synthesis machinery, therefore a clearer mechanistic understanding of the causal effectors downstream of mTOR would strengthen the manuscript. For example, the authors could use a non-drug resistant cell line (i.e. LIM1215) and overexpress eIF4E and/or eEF2, to see if upregulation of a specific step in translation can increase KRAS levels and drive drug resistance. expression is supported by data from only 2 cell lines. The authors should determine global protein synthesis rates in the panel of cell lines tested in Figure 5A, to see if the correlation between drug resistance and KRAS levels is a general trend.
3. In Figure 5, it would be beneficial if the authors could provide an example of a gene not dependent on codon usage, to see if its expression is affected by "translational upregulation". This would suggest that in the context of this cell line the effect of translational upregulation is specific to genes dependent on codon usage. 4. In Figure 6, it is not clear if treatment with translational inhibitors is killing the cells specifically through downregulation of KRAS, or through the inhibition of other targets. The authors should perform rescue experiments in the resistant cell lines by overexpressing HRAS, codon-optimized KRAS, and wild type KRAS. If the drugs kill through blocking the translational upregulation of KRAS, then expression of a HRAS or codon optimized KRAS, which are efficiently expressed, should rescue the killing effect of the translational inhibitors, whereas a wild type KRAS should not. Figure 6, it is not clear if "translational upregulation" is even occurring in these specific cell lines. An essential experiment would be to assay global protein synthesis rates in the parental and resistant cell lines. Minor comments:

For
1. Figure 1D is not interpretable in its current state. The data should be presented in a clearer way.

Reviewer #1 (Remarks to the Author):
Although this work is an extension of previous reports by these authors, in my opinion, the authors report important and translationally relevant implications of this biology that will be of interest to the oncology field in general and signal transduction researchers in particular. The observations appear to be robust and have been effectively described. I have no major concerns.
We thank Reviewer #1 for his/her review of our manuscript and appreciate his/her interest in the significance of our studies.

Reviewer #2 (Remarks to the Author):
In this manuscript, Ali et al. describe a role for codon usage in restraining KRAS expression as a barrier to drug resistance. The authors employ a pharmacological approach to show that overexpression of HRAS can drive drug resistance, while the closely related KRAS gene is less potent in driving drug resistance. The authors then show that expression of codon-optimized KRAS leads to higher protein levels, and confers drug resistance, similar to the effect observed when HRAS is overexpressed, implicating the codon usage of KRAS as a determinant of protein levels and drug resistance. Furthermore resistant cell lines with weakly activating KRASG12/13 mutations tend to have increased KRAS expression, and increased global protein synthesis rates. This is in contrast to the effect of strongly activating mutations such as KRASQ61R, which can confer drug resistance despite being expressed at low levels. Interestingly the authors note that cetuximab-resistant cell lines with weakly activating G12/13 mutations, as compared to the non-resistant parental line, exhibit increased KRAS expression. Importantly, cetuximab resistant cell lines are more sensitive than the parental lines to treatment with various inhibitors of protein synthesis. This effect is associated with a reduction in KRAS levels, suggesting that the drugs may kill at least in part through reducing KRAS expression.
Overall, the link between codon usage and RAS expression is established through welldesigned experiments in multiple cell lines, suggesting a general role for codon usage in regulating RAS expression. The observation that strongly-activating KRASQ61R mutations can drive resistance, despite low expression, while weakly-activating KRASG12/13 mutations require increased expression levels is very interesting, and suggests an novel mechanism of RAS-dependent drug resistance. However, the link between "translational upregulation" and increased KRAS levels is correlative. The manuscript would be greatly strengthened if the authors would provide evidence showing that the increased KRAS levels are the result of augmented protein synthesis, and are sufficient to confer drug resistance. Additionally, it is not clear if the lethality of drugs targeting the translational apparatus is due specifically to the decrease in KRAS levels, or due to either an overall decrease in protein synthesis or other targets affected by the compounds employed in the experiments. In principle, the manuscript is suitable for publication in Nature Communications, but these points should be addressed prior to acceptance.

In Figure 5, the authors show a correlation between global protein synthesis rates and KRAS expression, but it is not clear if the relationship between the two is causal. The observation that mTOR inhibitors abrogate KRAS expression implicates increased translation initiation and/or elongation rates in driving KRAS expression.
However, mTOR kinase phosphorylates many targets including components of the protein synthesis machinery, therefore a clearer mechanistic understanding of the causal effectors downstream of mTOR would strengthen the manuscript. For example, the authors could use a non-drug resistant cell line (i.e. LIM1215) and overexpress eIF4E and/or eEF2, to see if upregulation of a specific step in translation can increase KRAS levels and drive drug resistance.
We agree with Reviewer #2 that in addition to the correlation between global protein synthesis rates and KRAS expression, it would be beneficial to mechanistically define the effector downstream of mTOR that may be responsible for influencing KRAS expression. To address this point, we performed the experiment suggested by the reviewer, wherein we stably transduced a non-drug resistant cell line (LIM1215) with an empty vector construct or a construct expressing eIF4E, the key and limiting translational initiation component of the eIF4F complex. Overexpression of eIF4E caused an increase in KRAS protein levels (Supplementary Figure 5D). Through this experiment, we provide direct evidence that increasing this specific factor in translation initiation can increase KRAS expression. Interestingly, increased expression of KRAS, driven by eIF4E overexpression, was not sufficient to drive resistance to cetuximab in this non-drug resistant cell line (Supplementary Figure 5E). This latter result is likely due to the absence of a KRAS mutation in this cell line, which appears to be required for resistance. This observation is consistent with the finding that KRAS is frequently mutated in the setting of acquired cetuximab resistance. Figure 5, the link between global protein synthesis rate, cetuximab resistance, and KRAS expression is supported by data from only 2 cell lines. The authors should determine global protein synthesis rates in the panel of cell lines tested in Figure 5A, to see if the correlation between drug resistance and KRAS levels is a general trend.

In
We agree with Reviewer #2 that in the originally submitted manuscript, the link between global protein synthesis, cetuximab resistance, and KRAS protein expression was supported by data from only two cell lines, and that this should be expanded to additional cell lines. To determine global protein synthesis rates, puromycin labeling was performed in all six cell lines in Figure  5B. (We note that the data in Figure 5A was obtained from lysates that are available to us, however, we do not have all of these cell lines on hand. Thus, we performed western blotting and puromycin labeling on the subset of 6 of these lines that we have available.) We observed a statistically significant correlation between puromycin incorporation and KRAS protein expression in all six cell lines included in Figure 5B (Supplementary Figures 4A and 4B).

In Figure 5, it would be beneficial if the authors could provide an example of a gene not dependent on codon usage, to see if its expression is affected by "translational upregulation". This would suggest that in the context of this cell line the effect of translational upregulation is specific to genes dependent on codon usage.
To address this question, we transduced the panel of 8 KRAS mutant cell lines presented in Figure 5C with a FLAG-tagged NRAS construct. NRAS, a member of the RAS family of proteins, has an equal representation of rare and common codons, and therefore is not considered to be strongly reliant on codon usage. We found no correlation between the expression levels of ectopic NRAS protein and ectopic KRAS protein, suggesting that the increased protein translation we found to be transferrable to rare-codon containing constructs ( Figures 5D and 5E, Supplementary Figure 4C) is not transferrable to genes whose expression levels are less dependent on codon usage ( Supplementary Figures 4D and 4E). We have edited the manuscript text to accommodate this finding, which sheds important light on the translational upregulation observed in KRAS mutant lines by suggesting that this upregulation may be biased toward rare codon-enriched transcripts. Figure 6 This is an insightful question and one that we also were curious to answer. As per Reviewer #2's suggestions, we performed rescue experiments in our cetuximab-resistant clone to determine whether overexpression of common codon-containing constructs, specifically HRAS and codon-optimized KRAS (KRAS-COMMON, or KRAS-C), would be able to rescue the selective killing effect of the translational inhibitors used in our experiments. Using this assay, we were unable to rescue the phenotype of resistant cells (Panel A below), as evidenced by no change in the half-maximal growth inhibition values for 4EGI-1. Importantly however, although we were able to overexpress the common codon constructs in this experiment (Panel B), their expression levels diminished to background levels following treatment with translational inhibitors (Panel C). This latter finding suggests that in the presence of these inhibitors, overexpression of both endogenous and ectopic KRAS protein remains suppressed, obscuring our interpretation of these data. As an alternative approach, we performed KRAS knockdown experiments using short hairpin RNAs (shRNAs) in both our parental and resistant cells, and found that loss of KRAS expression significantly affected cell viability by prolonging cell doubling time only in the cetuximab-resistant derivative (Supplementary Figures 5F and 5G). Combined with our findings that cetuximab-resistant cells develop collateral sensitivities to translational inhibitors ( Figure 6D), which themselves suppress KRAS protein expression ( Figure 6E), these data suggest that the hypersensitivity of resistant cells to translational inhibitors parallels the effects of KRAS knockdown. Nevertheless, other alternative explanations are possible, which we note in the text.  Figure 6, it is not clear if "translational upregulation" is even occurring in these specific cell lines. An essential experiment would be to assay global protein synthesis rates in the parental and resistant cell lines.

For
This is an excellent point. To address this question, we performed puromycin labeling in the parental (P) and resistant clones (R1 and R2) presented in Figure 6A. Our data suggest that, as expected, the resistant clones demonstrate increased global protein translation (Supplementary Figure 5C). We agree with Reviewer #2 and believe that multiple independent experiments should be performed prior to making any conclusions. We would like to clarify that the conclusions based on western blot data were made following multiple experimental replicates (in most cases 2 and in some cases 3), and across multiple cell lines and/or resistant derivatives, with replicate or triplicate experiments leading to the same conclusions in all cases. Similarly, the experiments performed during the revision process were repeated twice in all cases. We apologize for not making this clear in the original manuscript and have now included the number of replicates for each experiment in the figure legends.

Throughout
Minor comments: 1. Figure 1D is not interpretable in its current state. The data should be presented in a clearer way.
In regards to this reviewer comment, we have increased the size of the font for each of the drugtreatment screens presented in Figure 5D. We hope that the general trend of HRAS G12Vmediated resistance scoring higher (higher overall enrichment scores) across the screens presented is now more clear to readers, and that these data help to underscore our observation that even when both HRAS G12V and KRAS G12V scored in any particular screen, the enrichment score for HRAS G12V -mediated resistance was always higher. However, if the reviewer has additional suggestions for how to more clearly present these data, we would certainly welcome them. Figure 6D, the authors switch between showing GI75 and GI50 values for different drugs. One or both values should be shown for all drug treatments.

In
Thank you for pointing out this oversight. In each case, we presented half-maximal or "true GI 50 " values, which corresponded to GI 75 or GI 50 values depending upon the drug used. To avoid confusion moving forward, we are now reporting GI 50 values as half-maximal growth inhibition values throughout.
Thank you again for considering our manuscript, and please let me know if we can provide any additional or clarifying information.