Identifying therapeutic targets for cancer among 2074 circulating proteins and risk of nine cancers

Abstract

Circulating proteins can reveal key pathways to cancer and identify therapeutic targets for cancer prevention. We investigate 2,074 circulating proteins and risk of nine common cancers (bladder, breast, endometrium, head and neck, lung, ovary, pancreas, kidney, and malignant non-melanoma) using cis protein Mendelian randomisation and colocalization. We conduct additional analyses to identify adverse side-effects of altering risk proteins and map cancer risk proteins to drug targets. Here we find 40 proteins associated with common cancers, such as PLAUR and risk of breast cancer [odds ratio per standard deviation increment: 2.27, 1.88-2.74], and with high-mortality cancers, such as CTRB1 and pancreatic cancer [0.79, 0.73-0.85]. We also identify potential adverse effects of protein-altering interventions to reduce cancer risk, such as hypertension. Additionally, we report 18 proteins associated with cancer risk that map to existing drugs and 15 that are not currently under clinical investigation. In sum, we identify protein-cancer links that improve our understanding of cancer aetiology. We also demonstrate that the wider consequence of any protein-altering intervention on well-being and morbidity is required to interpret any utility of proteins as potential future targets for therapeutic prevention.

Fig. 1. Manhattan Plot for the association of genetically predicted protein concentrations with cancer risk.

Association of genetically predicted protein concentrations with cancer risk estimated using Wald ratios presented as a Manhattan plot where position is given by cis-pQTL coordinate with a selection of cancer risk associations additionally labelled for their association with cancer risk and colocalization probability (PP4). Top result for each cancer endpoint provided. All tests are two-sided. Points highlighted as filled-in are those with PP4 > 0.7 with point size reflecting PP4 magnitude, which can vary between 0 and 1. Risk associations with MR p > 0.05 were not subject to colocalization analyses. Results labelled are those passing correction for multiple testing for the number of proteins analysed in this study per-cancer.

Fig. 2. Forrest plot for the associations of 40 protein and cancer risk and pan-cancer risk forrest plots for PLAUR and CTRB1.

Proteins associated with cancer risk (A) after correction for multiple testing for the number of proteins analysed in this study per-cancer. Odds ratio estimates are estimated using Wald ratios scaled per standard deviation increment in relative circulating protein concentrations. Confidence intervals (95% CI) are derived using the standard Wald ratio formula and reflect the precision of the cis-pQTL estimate in cancer GWAS scaled by the beta for the cis-pQTL association with protein concentrations. Sample sizes for cancer GWAS can be found in methods. Association of higher PLAUR with cancer risk (B) coloured by the colocalization probability (PP4), where MR Wald p < 0.05. Association for higher CTRB1 with cancer risk presented to demonstrate pancreas-specific association (C) with colour scheme as described above. All tests are two-sided.

Fig. 3. Potential consequences of genetically-predicted protein altering interventions as explemplified using FGFR3 and PDXK.

Potential consequences, presented as beta coefficients and confidence intervals (95% CI) of a protein-lowering intervention for two emblematic cancer risk proteins, FGFR3 and PDXK, that associated with higher risk of bladder and breast cancer, respectively, estimated using Wald ratios for traits where cis-pQTL had a p-value for association passing correction for multiple testing in main analyses. For reference, the predicted effect of lowering each protein by 1 SD is present for cancer risk at the top (italics and bold). Below these estimates we present the predicted effect of protein-lowering where colocalization and MR analyses suggest an aetiological link with other traits. Colocalization probability (PP4) is also presented for each trait. All tests are two-sided. Box colour indicates whether a predicted consequence may be beneficial (blue), harmful (red) or have uncertain consequence (grey) on health. Box size is proportionate to absolute beta from MR analyses while box opacity is proportionate to precision of this MR estimate.

Fig. 4. Stacked bar plot that depicts identified proteins associated with cancer by their current highest level of therapeutic investigation, which additionally has colours that stratified results by cancer site.

Figure displays a stacked bar chart, coloured by cancer endpoint for protein in question, mapping of cancer-risk proteins to currently available information on clinical/pharmaceutical investigations. These include cancer-risk proteins that are the target of a currently launched drug, under investigation at Phase I or II, under preclinical or biological testing, and those not known to be under current investigation.