Tumour-specific activation of a tumour-blood transport improves the diagnostic accuracy of blood tumour markers in mice

Abstract

Background: The accuracy of blood-based early tumour recognition is compromised by signal production at non-tumoral sites, low amount of signal produced by small tumours, and variable tumour production. Here we examined whether tumour-specific enhancement of vascular permeability by the particular tumour homing peptide, iRGD, which carries dual function of binding to integrin receptors overexpressed in the tumour vasculature and is known to promote extravasation via neuropilin-1 receptor upon site-specific cleavage, might be useful to improve blood-based tumour detection by inducing a yet unrecognised vice versa tumour-to-blood transport.

Methods: To detect an iRGD-induced tumour-to-blood transport, we examined the effect of intravenously injected iRGD on blood levels of α-fetoprotein (AFP) and autotaxin in several mouse models of hepatocellular carcinoma (HCC) or in mice with chronic liver injury without HCC, and on prostate-specific antigen (PSA) levels in mice with prostate cancer.

Findings: Intravenously injected iRGD rapidly and robustly elevated the blood levels of AFP in several mouse models of HCC, but not in mice with chronic liver injury. The effect was primarily seen in mice with small tumours and normal basal blood AFP levels, was attenuated by an anti-neuropilin-1 antibody, and depended on the concentration gradient between tumour and blood. iRGD treatment was also able to increase blood levels of autotaxin in HCC mice, and of PSA in mice with prostate cancer.

Interpretation: We conclude that iRGD induces a tumour-to-blood transport in a tumour-specific fashion that has potential of improving diagnosis of early stage cancer.

Funding: Deutsche Krebshilfe, DKTK, LOEWE-Frankfurt Cancer Institute.

Keywords: CEND-1; Early cancer detection; HCC; Tumour marker; iRGD; α fetoprotein.

Fig. 1

Intravenously injected iRGD increases blood AFP levels in HCC-bearing mice. (a) Experimental setup to study the effect of iRGD on blood AFP levels in Huh-7 xenografted nude mice. (b and c) Blood AFP levels before and after intravenous injection of iRGD (b, c, n = 12), RGD control peptide (c (n = 6)), and PBS (c (n = 6)) in mice with HepG2 xenografts. Human AFP was not detectable in the blood of mice without tumours. In (c) the fold increase of AFP with pre-injection level set to 1; lines and error bars represent geometric means and 95% CI. (d) Experimental setup to study the effect of iRGD on blood AFP levels in TGFα/c-myc HCC mice. (e and f) iRGD specifically increased the blood AFP levels in TGFα/c-myc HCC mice. TGFα/c-myc mice (20–24 weeks old) with HCC according to MRI or without HCC (f) were intravenously injected with iRGD (e, f), RGD control peptide (f) or PBS (f). Data are fold changes of blood AFP due to the treatments (n = 48: iRGD; RGD control peptide: n = 34; PBS: n = 15); lines and error bars indicate medians and 95% CI. Dashed line: upper 95% CI increase in blood AFP in the PBS-injected HCC mice. (g) AFP expression in HCCs of TGFα/c-myc mice. HCCs and liver tissues were excised from TGFα/c-myc mice. Pairs of the tissue lysates were analysed for AFP and β-actin content by immunoblotting. Band densities were measured densitometrically. The ratio of AFP/β-actin in the livers was set to 1. (h) iRGD-induced accumulation of Evans blue in HCCs in TGFα/c-myc mice with HCC. TGFα/c-myc mice with HCCs were co-injected with iRGD or PBS (n = 12 per group) and Evans blue (EB). The dye content of the tumours was related to that in the livers; lines and error bars indicate geometric means (b, c, and h) or medians (e and f) and 95% CI; dashed line: upper 95% CI of the measured EB content of a HCC from the PBS-injected animals. (b): paired t test for log-transformed data; (c) One-way ANOVA with multiple comparison post-hoc test for log-transformed data; (e): Wilcoxon signed-rank test for log-transformed data; (f) Kruskal–Wallis test with Dunn’s multiple comparison post-hoc test; (h): two-sample t test. The indicated fold increase in (b, e and h) is the ratio of the geometric means with 95% CI.

Fig. 2

iRGD increases the blood AFP concentration in mice with HCCs formed endogenously in fibrotic livers, but not in mice with liver fibrosis only. (a–c) iRGD had no effect on blood AFP in mice with liver fibrosis. Six-month-old Mdr2^−/− mice (a, n = 16–17) as well as mice treated with Ad-2D6 (b, n = 5–10) or CCl₄ (c, n = 3–5) for four weeks were bled before and after the injection of iRGD or RGD control peptide and was analysed for AFP content. (d–f) iRGD increased blood AFP in mice with liver fibrosis and HCC. (d) Experimental setup. (e and f) Mice with HCC were injected intravenously with iRGD (n = 19), RGD control peptide (n = 12), PBS (n = 9), or the mice were injected with iRGD prior to tumour development (n = 23). For the latter, mice treated for 12 weeks with DEN-CCl₄ received contrast-enhanced MRI to show absence of liver tumours at this time of induction of hepatocarcinogenesis; (f) fold increase of AFP with pre-injection level set to 1; dashed line: upper 95% CI increase in blood AFP in the PBS-, RGD control peptide-treated HCC animals and in iRGD-treated animals without tumour. (a–c, e, f): Lines and error bars represent geometric means (a–c) or medians (e and f) with 95% CI. (a–c): Unpaired t test for log-transformed data with Welch correction (b); (e): Wilcoxon matched-pairs signed-rank test; (f): Kruskal–Wallis test with Dunn’s multiple comparison post-hoc test. The indicated fold increase in (e) is the median of the ratios.

Fig. 3

iRGD-induced elevation of the blood AFP concentration depends on NRP-1 and the tumour blood concentration gradient of AFP. (a) Anti-NRP-1 prevented iRGD-induced increase in the blood AFP concentration. TGFα/c-myc tumour mice that displayed a robust iRGD-induced increase in blood AFP level one week earlier were injected with anti-NRP-1 and the effect of iRGD on blood AFP level was determined (n = 3 per group). Fold increase of AFP with pre-injection level at the first time point was set to 1. Lines and error bars represent medians and 95% CI. (b–e) iRGD-induced elevation of the blood AFP level correlated negatively with the pre-injection blood AFP level in TGFα/c-myc mice (b), DEN-CCl₄-HCC mice (c), mice with Huh-7 (d) or HepG2 xenografts (e). Spearman correlation r (b and c) and Pearson correlation r (d and e) with 95% CI, two tailed p-values and the log–log regression lines. (f) iRGD increased the blood AFP levels in HepG2 xenografted nude mice and low basal AFP (<67 ng/ml, n = 36, left panel), but not in animals with high basal AFP (>67 ng/ml, n = 12, right panel). Lines and error bars indicate geometric means and 95% CI. [(g) iRGD increased the blood AFP levels in TGFα/c-myc mice with HCC and normal basal AFP (<67 ng/ml, n = 36, left panel), but not in mice with elevated basal AFP (>67 ng, n = 12, right panel). Lines and error bars represent medians (left) or geometric means (right) with 95% CI. Significance was calculated with one sample t test (a, left) and the unpaired t test (a, right), paired t test (f and g, right) and Wilcoxon matched-pairs signed-rank test (g, left). The indicated fold increase in (f) is the geometric mean ratio with 95% CI. The indicated fold increase in (g) is the median of the ratios with 95% CI.

Fig. 4

iRGD induces an increase of the blood AFP concentration in mice with small HCCs and low basal blood AFP levels. (a) Basal AFP levels in DEN-CCl₄-treated mice prior to the tumour formation (week 15 of treatment) and in mice with small and larger HCCs. (b) iRGD increased blood AFP levels in DEN-CCl₄-treated mice with liver fibrosis and small HCCs (tumours with maximal diameters of 1–3 mm according to MRI) and low blood AFP levels (>500 ng/ml, n = 6), and in animals with larger HCCs (tumours with maximal diameters >3 mm according to MRI) and low (>500 ng/ml), but not in mice with small HCCs and high basal AFP. (c) DEN-CCl₄-treated mice with larger HCCs (>3 mm tumour diameter, n = 7) showed a stronger iRGD-induced increase of the blood AFP concentration than mice with smaller HCCs (<3 mm tumour diameter, n = 12). (d) Tumour size-dependence of the basal blood AFP level in TGFα/c-myc mice. AFP levels in TGFα/c-myc mice with no macroscopic liver tumours (6–7 weeks, n = 23, >20 weeks, n = 40) or tumours with maximal diameters of 1–3 mm (n = 13), 3–7 mm (n = 18) or >7 mm (n = 17) as graded by contrast-enhanced MRI. Dashed line: highest blood AFP concentration measured in a tumour-free animal. (e) iRGD responsiveness of blood AFP level in TGFα/c-myc mice with small (n = 13), medium (n = 18), or large liver tumours (n = 17). Fold increase of AFP with pre-injection level set to 1. (a–e) Lines and error bars represent geometric means (b, c and e top and left bottom) or medians (a, c, and e, right bottom) with 95% CI. (a) Kruskal–Wallis test with Dunn’s post-hoc multiple comparison; (b and e) paired t test or Wilcoxon matched-pairs signed-rank test (only e right, bottom) for log-transformed data; (c) Mann–Whitney U test. (d) One-way ANOVA for log-transformed data. (b and e) The indicated fold increase is the geometric mean ratio.

Fig. 5

iRGD induces an increase in autotaxin (ATX) levels in HepG2-xenografted mice, and of PSA in LNCaP-xenografted mice. (a) Blood was drawn from HepG2 xenografted nude mice (tumours 0.6–1.5 cm in diameter) 5 min before and 90 min after the i.v. injection of iRGD. The sera were analysed for ATX content (n = 4 per group); lines and error bars represent medians and 95% CI. (b and c) Blood was drawn from mice xenografted with LNCaP prostate cancers (tumours 0.8–1.3 cm in diameter) 5 min before and 90 min after the i.v. injection of iRGD (n = 11) (b) or control peptide (n = 5) (c) and analysed for PSA content. Fold increase of AFP with pre-injection level set to 1. Lines and error bars represent medians (a) or geometric means (b and c) with 95% CI. (a–c) Statistical significance was calculated with Kruskal–Wallis test with Dunn’s multiple comparison post-hoc test (a) and one sample t test (b and c). The indicated fold increase in (b) is the ratio of the geometric means with 95% CI. (d) Hypothetical model of the action of iRGD to induce a tumour-to-blood transport of tumour-secreted AFP, PSA and ATX. In the absence of iRGD, tumour-secreted AFP, PSA and ATX penetrate only slowly across the tumour endothelium (left panel), whereas in the presence of iRGD, the activation of NRP-1 by the CendR peptide generated from iRGD induces a paracellular and/or a transcellular transport across the vascular endothelium that transports AFP, PSA and ATX according to the tumour-blood concentration gradient (right panel). Created with BioRender.