In the present study, we investigated the levels of PCSK9, ANGPTL3 and Lp(a) in women with breast diseases of different levels of severity (benign, stage 0 and stage III cancers). We also investigated whether these lipid-related proteins were associated to the lipid profile or breast disease severity. The main finding of this study is that PCSK9 levels increase along with the severity of the breast disease, a finding that warrants confirmation in a larger cohort.
Severity of the breast disease versus the lipid panel
Although it is acknowledged that cholesterol is essential for tumor growth [61, 62], contradictory findings have been reported on the association between circulating lipoprotein levels and breast cancer risks [8, 11]. In the present cohort of women, none of the lipid parameters (total cholesterol, triglycerides, HDL-cholesterol, LDL-cholesterol, Apo B) was correlated to the severity of the breast disease (Table 3), a finding either explained by I) an insufficient power to detect changes in the lipid profile or II) a lack of association. Our study is in line with an analysis of the Women's Health Study that evaluated the prospective association of cancers (breast n = 864, colorectal n = 198, lung n = 190) with lipid profiles in 15,602 women. While lipids were associated with lung and colorectal cancer risks, no association was obtained between lipids and breast cancer risks . Nevertheless, a study published in 2008 by Shah et al including 125 women with breast cancer, 30 with benign lesions of the breast and 90 controls indicated that alterations in the lipid profile showed a significant correlation with breast cancer risks, disease status, and treatment outcome . The results of Shah et al were partly corroborated by a meta-analysis including fifteen prospective cohort studies involving 1,189,635 participants and 23,369 breast cancers. This meta-analysis suggested that triglycerides may be inversely associated with breast cancer risks while HDL-C might protect against breast cancer in postmenopausal women .
Clear evidence of a correlation between blood cholesterol levels and breast tumor histological features is still missing . Co-existing physiological conditions, such as an underlying metabolic syndrome, post-menopausal status or chemotherapy, can influence the levels of circulating lipids and consequently blur the association between lipid profile and breast cancer prognosis [63, 66–68].
ANGPTL3, Lp(a) and PCSK9 levels versus the lipid panel
No correlations were obtained between the the levels of (ANGPTL3, Lp(a) or PCSK9 and the lipid profile (Fig. 4A). The lack of association between Lp(a) levels and LDL particles is not an unexpected finding. Although apolipoprotein (a) is physically bound to LDL particles, Lp(a) concentrations are genetically determined  and are not correlated to LDL-C . The lack of correlation between ANGPTL3 and the lipid profile (Fig. 4A) is a finding in line with a Finnish study that found that ANGPTL3 serum protein concentrations did not predict lipid concentrations . Similarly, despite an increase in PCSK9 levels between benign disease vs stage III cancers (Fig. 3C), no changes in LDL-C, non-HDL or Apo B were observed (Tables 2 and 3). This observation was consistent with correlation studies between baseline PCSK9 levels and cholesterol measurements in 55 patients with different stages of lung cancer . While PCSK9 levels increased with cancer stage, cholesterol levels did not follow the same trend. We propose that either the rise of PCSK9 levels in our study was too small to lead to a significant increase in LDL-C levels, or that it did lead to a transient rise in LDL-C levels that was compensated by the uptake of LDL-C by cancer cells. Our study is not designed to explore these processes.
Breast disease severity versus ANGPTL3, Lp(a) and PCSK9 levels
Severity of breast cancer is assessed using a staging system based on how large the primary tumor is and how far it has spread within the body. Stages 0 to II tend to have a better long-term outcome compared to stages III to IV, which are qualified as “high-stage” cancers.
In our study, no statistical differences in ANGPTL3 or Lp(a) levels were observed between groups of increasing breast disease severity (Table 3, Figs. 1 and 2). This lack of association could be attributed to a low statistical power or a real absence of correlation with disease severity. The observation of a significant association between severity of breast disease and PCSK9 levels (Table 3, Figs. 3B and 3C) points toward a potentially important effect of PCSK9 in cancer, the mechanism of which deserves to be further investigated.
Our findings are in line with results reported in other human cancers. A recent in vivo study has shown that patients with gastric cancers had higher plasma levels of PCSK9 compared to age-matched healthy controls . Other studies have advocated that PCSK9 could be a prognostic marker for advanced non-small cell lung cancer  and for response to immune checkpoint inhibitors therapy . Combined with the present results, these studies support the need to confirm the association between PCSK9 elevation and enhanced cancer aggressivity.
Origin of circulating PCSK9
Our study was not designed to address the origin of the extra levels of PCSK9 measured in advanced breast cancers. A major organ responsible for PCSK9 levels in circulation is the liver, which is in charge of regulating cholesterolemia. A possibility would be that tumoral cholesterol uptake leads to a transient decrease in cholesterolemia which could trigger hepatic secretion of PCSK9. Higher PCSK9 levels would compensate tumoral cholesterol uptake and maintain cholesterolemia, without any apparent changes in the lipid profile, as observed in our study. Whether or not circulating levels of PCSK9 can partly originate from the tumor itself is an open question that needs to be addressed.
PCSK9 and tumor progression
A recent study by Suh et al showed that nucleoprotein Ahnak mediated B16F10 melanoma cells mestatasis in mice via enhanced PCSK9 expression . Actual data tend to support a pathogenic and pro-tumoral role for PCSK9, by both maintaining cholesterol supplies and by promoting MHC-I degradation on tumor cells, favoring immune escape of tumors . Whether the increase in PCSK9 plays an active role in the progression of breast cancer requires further investigation. Comparing tumoral levels of PCSK9, MHC-I and HLA-C using standard methods (ex: ELISA) in tumors of increasing disease severity could help decipher the role of PCSK9 in tumors.
The present study was performed on a limited amount of plasma samples (fourteen at the benign stage, nine at the premalignant stage and twenty-three stage III breast cancers). Due to the limited availability of samples, the control group consisted of women having a benign breast disease grouped with samples from women with a stage 0 breast cancer. While such a merge of non-cancerous samples with early-stage cancerous samples to form a control group may have blurred associations between PCSK9 and breast disease severity (Fig. 1C), a statistically significant association was obtained once samples were stratified according to breast disease severity (Figs. 3B and Table 3). A significant difference in HDL-cholesterol (1.3 mmol/L vs 1.8 mmol/L, p = 0.01) was observed between the benign disease of the breast group and the stage 0 breast cancers (Table 1). However, HDL is not associated with PCSK9 (Fig. 4A) and therefore this difference is not likely to invalidate our findings. HDL values in all groups are in the normal range for women (1.3 mmol/L and above). The lack of association between breast cancer stage and levels of ANGPTL3 or Lp(a) could be due to lack of statistical power. Despite a low number of samples, a statistically significant increase in PCSK9 levels was detected with increased severity of breast disease. If our finding is confirmed in a larger study, PCSK9 increase in breast cancer is likely to be an important mechanism of progression (cholesterol supplies, immune escape through tumoral MHC-I degradation) detectable even in our small cohort. Information on medications such as statins was unknown, which is a limitation since statins may change levels of both PCSK9 and circulating lipid levels. Finally, ethnical origins of women were not documented.