Although variable, general staining was stronger for cytoplasmic RRM1. Several clients confirmed weak or adverse staining.Fig three. Barplots for relative RRM1 and RRM2 protein expression ranges in cytoplasm (1462249-75-7 built-in depth, scale from to 3) by (A) histology and (B) quality. The two RRM1 and RRM2 expression ended up 940310-85-0 increased in squamous carcinoma than adenocarcinoma (p < 10). RRM1 expression did not correlate with tumor grade (rho = 0.04, p = 0.55) while RRM2 expression correlated positively with grade (rho = 0.44, p < 10 for cytoplasm and rho = 0.36, p < 10 for nucleus).Fig 4. Kaplan Meier curves for RRM1 and RRM2 in all patients. (A) cytoplasmic RRM2, split by median expression levels (p = 0.002, hazard ratio = 1.70). (B) cytoplasmic RRM1, split by median expression levels (p = 0.497, hazard ratio = 1.12).smoking less than one month prior to diagnosis), RRM2 was not predictive for survival (p = 0.85, hazard ratio = 1.13). However, in non-smokers, high RRM2 expression was associated with worse survival (p < 10, hazard ratio = 54.29). In former smokers, high RRM2 expression was also associated with poor survival (p = 0.002, hazard ratio = 5.63). On further subdividing former smokers, RRM2 was a significant predictive variable for survival only in those patients quitting over 10 years prior to diagnosis (p = 0.001, hazard ratio = 12.75). In those quitting less than 10 years, p = 0.25 and hazard ratio = 2.56 by the univariate Cox model. In women (both older and younger age groups), RRM2 also significantly predicted worse survival (p = 0.0002 and hazard ratio = 5.11) but this was not the case in men (p = 0.413 and hazard ratio = 1.51). When examining survival related to nuclear expression of RRM2, findings were similar except not as strong (Table 2). RRM1 was not a significant survival predictor in any of these patient subgroups. Kaplan-Meier curves for cytoplasmic RRM2 in these subgroups using the median value cut point are shown in Fig 5. Kaplan-Meier curves showing positive (staining seen in any cell) vs negative (no detectable staining) are shown in Fig 6. Table 2 shows predictive value (p-values and hazard ratios) for cytoplasmic RRM1, cytoplasmic RRM2 and nuclear RRM2 as continuous variables in a univariate Cox model. Cytoplasmic RRM2 remained an independent predictor of survival in the multivariate Cox model, taking into account stage, age and grade (Tables 3 and 4). Combining expression of RRM1 with RRM2 did not predict patient outcomes better than RRM2 alone. S1 Fig illustrates this with representative Kaplan-Meier curves for the different subgroups.Several studies have previously been performed on RRM1, RRM2 and RRM2b expression and outcomes in NSCLC. However, most studies examined the NSCLC population as a whole, without separating by gender and/or smoking status, for which differences in biological properties have been noted [5, 34]. In this paper, we reported immunohistochemical studies of RRM1 and RRM2 on a NSCLC tissue microarray. The stronger staining in cytoplasmic vs. nuclear compartments is consistent with previous studies of RNR being a constitutively cytosolic enzyme [9, 35]. High RRM2 levels (mainly cytoplasmic) predicted significantly worse overall survival in women, non-smokers and former smokers who had quit smoking at least 10 years prior. RRM2 was not a significant predictor of survival in men, current smokers and former smokers who had quit smoking more recently. In contrast, RRM1 was neither predictive of survival in the entire patient population nor any patient subgroup and the combination of RRM1 and RRM2 expression together did not add to the informative value of RRM2 alone.