In silico analysis of overall survival with YBX1 in male and female solid tumours

The Y-box binding protein-1 (YBX1) gene codes for a multifunctional oncoprotein that is increasingly being linked to the regulations of many aspects of cancer cell biology. Disparities in treatment outcomes between male and female cancer patients are increasingly reported. This study aimed to examine the relationship between YBX1 expression and overall survival in male and female patients with solid tumours. Overall survival and YBX1 expression data for cohorts of male and female cancer patients obtained from freely available databases were analysed with a cox proportional hazard model with covariates of biological sex and YBX1 expression. Kaplan–Meier curves and Violin plots were constructed for segregated male and female cohorts. High YBX1 expression was significantly associated with poor survival in 2 female-only and 4 mixed-sex cancer sites. In female lung cancer patients, better survival and lower YBX1 expression were identified. The clinical importance of YBX1 expression in cancer ought to be evaluated in a sex-specific manner, especially in lung cancer.


Co-expression analysis
Co-expression coefficients were generated by downloading data from the "Coexpression tab" in the cBioportal web service.The list of genes located on chromosome X was downloaded from Uniprot and validated in Human Genome Organisation (HUGO) database.The chromosome X genes that showed a Spearman'S correlation coefficient greater or lower than 0.25 and − 0.25, respectively, and q value < 0.05 (FDR < 0.5) were deemed correlated with YB-1.

Ethics declaration
All data generated or analysed during this study was downloaded from the freely accessible databases outlined in Table 1.This data was irrevocably anonymous and is deposited on these open access platforms.All methods were carried out in accordance with relevant guidelines and regulations.

High YBX1 expression is associated with reduced overall survival
We first analysed available survival data in all 15 identified cancer cohorts using a Cox proportional hazard model.Application of the Benjamini-Hochberg procedure to keep the false discovery rate at α = 0.05 for all cohorts yielded a threshold significance value of α bh = 0.02 .A significant relationship between YBX1 expres- sion and survival was detected in 6 cancer types: breast, liver, lung, renal papilloma, uterine cancer, and sarcoma (Table 2).In all these sites, the hazard ratio for YBX1 was > 1, indicating that higher expression levels were associated with poorer survival (Table 2 and Fig. 1).As expected, biological sex did not affect survival in female only cancers (Breast, uterine cancer).In cancer sites affecting both males and females, biological sex did not interact with YBX1 expression and did not affect survival in this analysis (Hazard ratio = 1) (data not shown).

YBX1 expression and biological sex
We next focused on the 4 cancer sites that affect both males and females, where YBX1 expression was identified to significantly affect survival: liver, lung, renal papilloma, and sarcoma.First, we constructed Kaplan-Meier curves to compare the survival of segregated male and female patient cohorts (Fig. 2).Lung was the only cancer type displaying a highly significant difference in survival when the data was analysed according to sex.Second, we examined the distributions of YBX1 expression levels in both male and female cohorts and confirmed that these were approximately normal, with two-tailed t-tests.Finally, we compared YBX1 expression between sexes (Table 3).A Šidák variation of the Bonferroni correction yielded a threshold of α bh = 0.0127 .At this threshold, sex differences in YBX1 expression for the Lung cancer cohort were highly significant with a Cohen's D of 0.363, indicating a medium to large effect size.A violin plot of the YBX1 expression distribution indicates higher expression in males, compared to females in lung cancer (Fig. 3).
Table 2. YBX1 expression and biological sex survival analysis in 15 cancer cohorts.*Hazard ratios are given per 1000 units of gene expression Quoted p-values and hazard ratios refer to YB-1 expression.Direct sex effects did not reach significance threshold and are not included here.

YBX1 expression and the X chromosome
We next focused on one cancer site that is known to affect both males and females differently [35][36][37][38] , where YBX1 expression was not identified to significantly affect survival in our analysis: bladder cancer.We generated correlation coefficients for the expression of YBX1 and individual X-linked genes in both male and female patients.In total 47 (male) and 115 (female) chromosome X genes were identified to co-express positively or negatively with YBX1 (Fig. 4).N = 37 were common to both sexes (Supplementary material S2).Of those, DKC1 held the highest positive correlation coefficient (0.36) and VGLL1 the lowest (− 0.42).Kaplan-Meier analysis identified an association between expression and overall survival in both male and female cohort for VGLL1 but not DKC1 (data not shown).Of the 78 genes uniquely identified in the female cohort (Supplementary material S2), VBP1 held the highest positive correlation coefficient (0.4) and FOXO4 the lowest (− 0.40).In these female patients, low VBP1 was associated with poorer overall survival (HR = 1.87 (1.05-3.01),p = 0.03).No association was detected for FOXO4.In the male cohort (Supplementary material S2), all but MOSPD1 appeared associated with overall survival on Kaplan-Meier analysis (data not shown).

Discussion
Sex is a fundamental biological variable increasingly studied as a factor influencing cancer treatment response 39 .Cancer affects men and women 40 ; but we treat patients.This sex-neutral approach results from the belief that circulating sex hormones dominate sexual differentiation biology 41 and the practice of sex data pooling 39 .Our efforts, however, yield unequal success between the sexes 42 .The predicted rise in the 19.3 million annual new cancer cases 40 will worsen the clinical and societal impact of treatment resistance and innovation in cancer management is a clinical priority.YB-1 is a multifunctional protein involved in both the transcriptional and translational regulation of gene expression 43 .The detection of this oncoprotein in tumour specimens is increasingly linked to poor patient outcomes.But the importance of YBX1 gene expression remains poorly documented.In the Prognoscan database 44 , YBX1 expression is associated with an increased hazard ratio for overall survival with Breast cancer, Lung cancer and prostate cancer (data not shown).We used available data for 15 cancer types to examine the link between YBX1 expression and survival outcomes.Our analysis identifies that high YBX1 expression is associated with poor survival in 6 cancer types.
YB-1 controls almost all DNA and mRNA dependent processes in the cell such as cellular differentiation, proliferation and stress response 43 .The regulation of these critical processes is increasingly linked to biological sex.This fundamental biological variable is defined by the presence of genetic information provided by the X and Y chromosomes, whose regulation and loss are proving relevant to cancer biology and treatment Figure 1.Log of the Hazard ratio against YB-1 expression levels for cancers with significant expression effects in Table 2.The shaded region depicts the 95% confidence interval.Note the varying axes limits for both log hazard ratio and expression level.outcomes [7][8][9]36,38,[45][46][47] . Lack of sexanalysis in preclinical and interventional studies was proposed to increase the risk for an effect being lost or claimed where it only applies to one sex 16 .In biomedical research analysis of the literature identified the underrepresentation of female animals and a lack of sex-specific reporting 48 .Our analysis expands earlier report that a correlation between the expression of YB-1 and X-linked genes exists [7][8][9] .In both male and female patients with bladder cancer, we identified 37 interactions common to both sexes, 10 limited to male patients and 78 to female patients.The relevance of biological sex in this disease is increasingly reported and could affect disease classification, and immune responses 36,38,47 .Further characterisation of the relevance of X-linked genes to the behaviour of malignant diseases is warranted.This study aimed to determine whether the relationship between YBX1 expression and overall survival is affected by the biological sex categorisation of the patient cohorts 49 . Co proportional hazard analysis of available data failed to identify biological sex as a co-variate significantly affecting survival in all 15 cancer sites examined.Similarly, meta-analysis of YB-1 protein expression, survival and clinicopathological features indicated that overexpression correlates with worse overall survival, but no association was identified with sex on multi-variate analysis 13 .Yet, Kaplan Meier curves were significantly different between male and female lung cancer patients.In lung cancer, the analysis of gene expression signatures according to the sex of the patients included revealed distinct cluster groups 50 .Our analysis of YBX1 expression identified a significant difference between the expression distributions of the male and the female cohorts in the case of lung cancer, which might be related to the stark differences in mortality between sexes.While we were unable to find a suitable data set for male-specific disease like prostate cancer, the prognoscan database 44 suggests that YBX1 expression increases hazard ratio in prostate cancer survival, and future work is needed to elucidate why this might be the case.
This work serves to highlight that the generation of sex-based analysis could refine the relevance of candidate genetic markers and emerging therapeutic targets.Further evaluation of the biological and clinical implications of our findings is needed.Future studies aimed as assessing he biological functions and clinical importance of YBX1, and its protein product in cancer ought to consider the biological sex of their models and patients, especially in lung cancer.This could be of particular relevance to the development of senolytic drugs, such as the YB-1 inhibitor fisetin, for the treatment of cancer 51 .In lung cancer, several reports already indicate the capacity of this drug to affect lung cancer cell growth, migration and apoptosis [52][53][54] , but unfortunately this effect was only tested in male lung cancer models.

Figure 2 .
Figure 2. Kaplan-Meier survival curves for male and female patients cohorts in lung, renal papilloma, liver and sarcoma.The p-values for the differences between male and female cohort survival is given in the figure for each cancer type.

Figure 3 .
Figure 3. Sex differences in YBX1 expression distribution for Lung cancer.The red dot indicates distribution medians, and the diamond indicates distribution means.

Figure 4 .
Figure 4. Waterfall plots of the X-linked genes identified as associated with YB-1 on male and female bladder cancer patients.

Table 1 .
Cancer cohorts and sources.