Cardiac factors may mediate the association between HF and cancer
Heart failure stimulates tumor growth by circulating factorsLiterature - Meijers WC, Maglione M, Bakker SJL et al. - Circulation 2018; 138:678-691
Introduction and methods
HF as a consequence of cancer treatment has been widely studied and is an accepted phenomenon. However, cancer development as a consequence of HF has not been studied. A few recent studies have shown a relationship between HF and incidence of cancer [1-4], but causality has not been demonstrated. Therefore, this study examined the possibly causal relationship between HF and cancer development.
Two mouse models were used with mice highly susceptible to spontaneous intestinal adenoma development. In the first model, myocardial infarction (MI) was induced in mice (n=22) and results were compared to sham-operated control mice (n=10). In the second model, mice received an MI heart (n=17) or a sham donor heart (n=7) while the native heart was left in situ. Myocardial expression levels of five selected candidate proteins with predictive value were measured in controls and HF patients using samples from the PREVEND study (Prevention of Renal and Vascular End-Stage Disease) and the VitD-CHF study (Study to Investigate the Effects of Vitamin D Administration on Plasma Renin Activity in Patients With Stable Chronic Heart Failure). Furthermore, the relationship between cardiac and inflammatory biomarkers and new-onset cancer was studied in 8319 subjects enrolled in the PREVEND study.
- In mice with HF compared with sham-operated mice, significantly more (57 vs 34, P<0.001) and larger (1.69 mm vs 1.44 mm, P<0.001) polyps were observed. A 2.4-fold increase in tumor load was seen in HF mice (P<0.0001). Polyp bleeding, a well-described feature of this model, resulted in anemia, splenomegaly and significantly higher spleen weight in HF compared with sham mice (P=0.038).
- Fibrosis was positively associated with tumor load (β=0.51, P=0.016) and LVEF was negatively associated with tumor load (β=−0.63, P=0.002).
- Similar results were observed for mice with heart transplants: significantly more (55 vs 39, P<0.01) and larger (1.97 mm vs 1.51 mm, P<0.01) polyps, 2.4-fold increased tumor load (P<0.0001) and a significantly larger spleen (1.4-fold increase with P<0.05) in mice with an MI heart transplant compared to mice receiving a sham-operated heart transplant. Again, tumor load and fibrosis (β=0.89, P<0.01) and tumor load and LVEF ((β=−0.60, P=0.002) were associated.
- A literature search on candidates of secreted cardiac markers resulted in the identification of: α-1-antitrypsin (SerpinA1), α-1-antichymotrypsin (SerpinA3), fibronectin, ceruloplasmin, and paraoxonase 1.
- RNA levels of all 5 genes were upregulated in the left ventricular tissue of mice with HF compared with sham in the first model (P<0.05) and in the transplant model 3 out of 5 (SerpinA3, fibronectin, and paraoxonase 1).
- Plasma levels of all 5 proteins were 30% to 100% upregulated in patients with chronic HF enrolled in the VitD-CHF study (n=101) compared to healthy subjects enrolled in the PREVEND study (n=180).
- In 8319 subjects of the PREVEND study, subjects in the tertiles 2 and 3 of NT-proBNP, an established biomarker of HF, had an increased risk of developing all-cause cancer and colorectal cancer (both P<0.0001).
- Cardiac (NT-proBNP, high-sensitivity troponin T, MR-pro-atrial natriuretic peptide) and inflammatory biomarkers (C-terminal proendothelin-1, MR-pro-adrenomedullin, high-sensitivity C-reactive protein) were associated with new-onset cancer, also after adjustment for known risk factors for cancer.
Presence of HF is associated with increased formation and accelerated tumor growth in a mouse model of intestinal cancer. Several myocardial biomarkers were identified that were upregulated in mice with HF and also in patients with chronic HF. In a cohort study, cardiac markers, including NT-proBNP, and inflammatory markers, including C-reactive protein, were associated with the prediction of new-onset cancer. Altogether, these results suggest that HF may be a risk factor for incident cancer.
In their editorial article, Kitsis, Riquelme and Lavendero  raise several questions after summarizing the mouse experiments by Meijers et al.: ‘Can these observations be extrapolated to other precancerous or cancerous lesions? Would heart failure of nonischemic origin elicit the same response? Does heart failure also promote metastasis?’ Although they recognize the novelty of a possible molecular mechanism, they mention that a more comprehensive assessment is needed to definitively link heart disease and cancer. Moreover, they believe an unbiased proteomics approach would be more appropriate to identify the biomarkers connecting HF and tumor growth and experiments with altered levels of these mediators are needed to prove cause and effect. They end by saying that these are ‘potentially groundbreaking results that will stimulate further delineation of the connections between heart disease and cancer’ and ’we may be at the gates of a new scientific research field’.