Angiotensin inhibition enhances drug delivery and potentiates chemotherapy by decompressing tumour blood vessels

Abstract

Cancer and stromal cells actively exert physical forces (solid stress) to compress tumour blood vessels, thus reducing vascular perfusion. Tumour interstitial matrix also contributes to solid stress, with hyaluronan implicated as the primary matrix molecule responsible for vessel compression because of its swelling behaviour. Here we show, unexpectedly, that hyaluronan compresses vessels only in collagen-rich tumours, suggesting that collagen and hyaluronan together are critical targets for decompressing tumour vessels. We demonstrate that the angiotensin inhibitor losartan reduces stromal collagen and hyaluronan production, associated with decreased expression of profibrotic signals TGF-β1, CCN2 and ET-1, downstream of angiotensin-II-receptor-1 inhibition. Consequently, losartan reduces solid stress in tumours resulting in increased vascular perfusion. Through this physical mechanism, losartan improves drug and oxygen delivery to tumours, thereby potentiating chemotherapy and reducing hypoxia in breast and pancreatic cancer models. Thus, angiotensin inhibitors -inexpensive drugs with decades of safe use - could be rapidly repurposed as cancer therapeutics.

Figure 1. Collagen and hyaluronan interact to compress tumour blood vessels.

(a) Representative image from intravital multiphoton microscopy of perfused tumour vessels (green) and collagen (blue), showing that high collagen levels colocalize with low perfusion in an E0771 breast tumour. Scale bar, 200 μm. (b) Histology images of vascular perfusion in orthotopic AK4.4 pancreatic tumours with high versus low collagen levels. High local collagen I levels (blue) appear to colocalize with collapsed vessels (red, collapsed; green/yellow, perfused) in vivo. Scale bar, 100 μm. (c) Correlation of perfused vessel fraction versus tumour matrix area fractions in multiple orthotopic pancreatic tumour models (AK4.4 and L3.6pl) in mice. Following lectin injection and animal killing, perfusion was quantified as the fraction of vessels that are both lectin- and CD31-positive out of all CD31-positive vessels. Perfusion inversely correlates with both hyaluronan (R=−0.79, P<0.001, Pearson’s correlation) and collagen I (R=−0.78, P<0.001, Pearson’s correlation), but has a stronger inverse correlation (R=−0.86, P<0.001, Pearson’s correlation) with the average matrix area fraction. (d) Grouping these tumours into those with either low (<17%) or high (≥17%) collagen reveals that perfusion does not correlate (R=−0.33, Pearson’s correlation) with hyaluronan in collagen-poor tumours but does inversely correlate (R=−0.71, P=0.004, Pearson’s correlation) in collagen-rich tumours. (e) In contrast, grouping the tumours into those with either low (<33%) or high (≥33%) hyaluronan shows that perfusion inversely correlates with collagen I in both hyaluronan-poor (R=−0.73, P=0.003, Pearson’s correlation) and hyaluronan-rich (R=−0.57, P=0.040, Pearson’s correlation) tumours.

Figure 2. Losartan decreases profibrotic stromal activity in tumours.

(a) Expression levels of fibrosis-related genes in murine cancer-associated fibroblasts (CAFs) isolated from orthotopic AK4.4 pancreatic tumours. AK4.4 tumours were orthotopically implanted in αSMA^P-dsRed/FVB mice, resulting in αSMA+ CAFs that express dsRed. These mice were treated with losartan or saline (control), then the CAFs from their tumours were isolated using fluorescence-activated cell sorting for dsRed. The isolated control CAFs express high mRNA levels of collagen I (Col1), hyaluronan synthases (HAS)1–2 (Has1, Has2), TGF-β1 (Tgfb1) and CCN2 (Ctgf), with low levels of HAS 3 (Has3) and moderate levels of ET-1 (Edn1). Levels normalized to GAPDH expression. (b) Comparison of expression levels of fibrosis-related genes in the CAFs isolated from losartan- or saline-treated (control) mice. Losartan reduces the mRNA expression of each of these profibrotic genes in CAFs (P<0.05, Student’s t-test). Levels normalized to saline control. (c) Histology images showing the effect of losartan on tumour TGF-β1 expression. Scale bar, 100 μm. (d) Immunohistochemical analysis of TGF-β1 expression with losartan treatment. Losartan reduces the expression of TGF-β1 in E0771 (*P=0.048, Student’s t-test) and AK4.4 tumours (**P=0.044, Student’s t-test). (e) Histology images showing the effect of losartan on tumour CCN2 expression. Scale bar, 100 μm. (f) Immunohistochemical analysis of CCN2 expression with losartan treatment. Losartan reduces the expression of CCN2 in E0771 (*P=0.044, Student’s t-test) and AK4.4 (**P=0.046, Student’s t-test) tumours. Scale bar, 100 μm. Animal numbers n=4 (CAF expression), n=7–8 (E0771 TGF-β1, CCN2), n=4–6 (AK4.4 TGF-β1, CCN2). Error bars indicate s.e.m.

Figure 3. Losartan reduces matrix and stromal density in tumours.

(a) Histology images showing the effect of losartan on tumour collagen levels and perfusion. Scale bar, 100 μm. (b) Immunohistochemical analysis of tumour collagen levels following losartan treatment. Losartan decreases the collagen I-positive area fraction in E0771 (*P=0.040, Student’s t-test) and AK4.4 (**P=0.022, Student’s t-test) tumours. (c) Histology images showing the effect of losartan on tumour hyaluronan levels. Scale bar, 100 μm. (d) Immunohistochemical analysis of tumour hyaluronan levels following losartan treatment. Losartan also reduces the hyaluronan-positive area fraction in E0771 (*P=0.048, Student’s t-test) and AK4.4 (**P=0.019, Student’s t-test) tumours, as assessed using a hyaluronan-binding protein probe. (e) Histology images showing the effect of losartan on tumour αSMA+ CAF levels. Scale bar, 100 μm. (f) Immunohistochemical analysis of αSMA+ CAF density with losartan treatment. Losartan reduces the CAF density in E0771 (*P=0.019, Student’s t-test) and AK4.4 (**P=0.040, Student’s t-test) tumours. Animal numbers n=5–7 (E0771 collagen), n=4–6 (AK4.4 collagen), n=4–5 (E0771 hyaluronan), n=4 (AK4.4 hyaluronan), n=5 (E0771 αSMA), n=5–6 (AK4.4 αSMA). Error bars indicate s.e.m.

Figure 4. Stromal angiotensin signalling induces matrix production in tumours.

(a) Tumour collagen and (b) hyaluronan levels in angiotensin-II-receptor-1 (AT1)-knockout mice. Orthotopic E0771 breast tumours implanted in AT1-knockout mice (Agtr1a^−/−) have (a) a lower collagen I area fraction (*P=0.026, Student’s t-test) and (b) a lower hyaluronan area fraction (*P=0.011, Student’s t-test) than E0771 tumours implanted in wild-type C57BL/6 mice. (c) Tumour collagen and (d) hyaluronan levels in angiotensin-II-receptor-2 (AT2) knockout mice. Orthotopic E0771 breast tumours implanted in AT2-knockout mice (Agtr2^−/−) have (c) a higher collagen I area fraction (*P=0.047, Student’s t-test) and (d) a higher hyaluronan area fraction (*P=0.012, Student’s t-test) than E0771 tumours implanted in wild-type C57BL/6 mice. Tumours were time- and size-matched at ~200 mm³. (e) Expression of AT1 (Agtr1a and Agtr1b genes) and AT2 (Agtr2 gene) mRNA in murine cancer-associated fibroblasts (CAFs) and cancer cells. CAFs isolated from AK4.4 tumours express over one order of magnitude more AT1 than either E0771 or AK4.4 cancer cells. Similarly, the CAFs express over two orders of magnitude more AT2 than either E0771 or AK4.4 cells. (f) Expression of AT1 (red) in αSMA+ CAFs (green) and cancer cells in tumours. We imaged CAFs based on αSMA+ expression in E0771 tumours. We found that some CAFs in these tumours express AT1 at high levels, whereas other cells express AT1 at low levels. Colocalization is shown in yellow. Scale bar, 25 μm. (g) Expression of AT2 (red) in αSMA+ CAFs (green) and cancer cells in tumours. Most CAFs express high levels of AT2, whereas some other cells express similar AT2 levels. Colocalization is shown in yellow. Scale bar, 25 μm. Animal numbers n=4–5. Error bars indicate s.e.m.

Figure 5. Losartan targets solid stress in tumours.

(a) Solid stress levels in tumours after angiotensin inhibition using losartan. Solid stress was assessed using an ex vivo technique involving the measurement of the extent of tumour tissue relaxation (tumour opening relative to tumour diameter) following a stress-releasing incision, with larger openings indicating higher stress. Through its antimatrix effects, losartan reduces solid stress in E0771 (*P=0.049, Student’s t-test) and AK4.4 (**P=0.043, Student’s t-test). (b) Losartan reduces solid stress in additional models, including 4T1 breast tumours (*P=0.036, Student’s t-test) and Pan-02 pancreatic tumours (**P=0.0092, Student’s t-test). Animal numbers n=5 (E0771), n=8–9 (AK4.4), n=10–11 (4T1), n=4–8 (Pan-02). Error bars indicate s.e.m.

Figure 6. Losartan decompresses tumour vessels to increase drug and oxygen delivery.

(a) Representative images from intravital multiphoton microscopy of perfused tumour vessels (green) and collagen (blue), showing that losartan increases the density of perfused vessels in an E0771 breast tumour. Scale bar, 1 mm. (b) Perfused vessel fractions after angiotensin inhibition using losartan. Losartan increases the fraction of vessels that are perfused in orthotopic E0771 breast (*P=0.038, Student’s t-test) and AK4.4 pancreatic (**P=0.039, Student’s t-test) tumours. (c) Losartan also increases the fraction of vessels with open lumen in E0771 (*P=0.040, Student’s t-test) and AK4.4 (**P=0.015, Student’s t-test) tumours, indicating decompression as the mechanism. (d) Vessel density and (e) vessel length following angiotensin inhibition using losartan. Losartan does not affect vessel density, as quantified by the vessel number density (d) and the total vessel length (e), indicating no antiangiogenic effect at this 40 mg kg⁻¹ dose. (f) Representative images from intravital optical frequency-domain imaging of perfused vessels with losartan treatment. E0771 tumours in control mice have a low density and poor distribution of perfused vessels in three dimensions, whereas losartan-treated mice showed a more even distribution and higher density of perfused vessels. Scale bar, 1 mm. (g) Small-molecule drug delivery to tumours and various organs after angiotensin inhibition with losartan. Losartan increases the accumulation of the small-molecule chemotherapeutic 5-FU in AK4.4 pancreatic tumours by 74% (*P=0.0063, Student’s t-test), while not affecting accumulation in the normal organs. (h) Oxygen delivery to tumours measured by phosphorescence quenching microscopy during angiotensin inhibition using losartan, with (i) representative images. Losartan maintains the level of oxygenation (h) in the tissue, versus control tumours that become progressively more hypoxic with time (*P=0.030, Student’s t-test) as the tumours grow from 3 to 5 mm in diameter. Losartan increases oxygenation in some tumours (i) whereas all control tumours decrease in oxygen levels. Losartan also appears to result in a more homogenous distribution of well-oxygenated tumour tissue. Scale bar, 100 μm. (j) Hypoxic fraction in tumours measured by pimonidazole injection and staining following angiotensin inhibition with losartan. Losartan decreases the hypoxic fraction in E0771 tumours (*P=0.019, Student’s t-test) because of the increase in oxygen delivery. Animal numbers n=7–9 (vessels), n=4–5 (vessel lumen), n=4 (drug delivery), n=6 (oxygen delivery), n=6–7 (hypoxia). Error bars indicate s.e.m.

Figure 7. Losartan potentiates chemotherapy.

(a) Quantification of tumour growth rates, based on the time to reach double the initial volume, for orthotopic E0771 breast tumours in response to treatment with losartan or saline control (40 mg kg⁻¹ daily from day 0 onwards) in combination with either the small-molecule chemotherapeutic doxorubicin or saline control (2 mg kg⁻¹ every 3 days from day 1 onwards). Doxorubicin and losartan monotherapy induce no significant growth delay versus the control treatment in these aggressive tumours. In contrast, their combination greatly limits tumour growth (*P=0.040, Student’s t-test). (b) Quantification of tumour growth rates, based on the time to reach double the initial volume, for orthotopic 4T1 breast tumours using the same treatments as with E0771. Doxorubicin and losartan monotherapy induce no significant growth delay versus the control treatment. In contrast, their combination greatly limits tumour growth (*P=0.024, Student’s t-test). (c) Volumes of orthotopic AK4.4 pancreatic tumours on day 7 in response to treatment with losartan or saline control (40 mg kg⁻¹ daily from day 0 to 7) in combination with either the small-molecule chemotherapeutic 5-FU or saline control (60 mg kg⁻¹ on days 2 and 6). 5-FU and losartan monotherapy induce no significant growth delay versus the control treatment, whereas their combination greatly inhibited tumour growth (*P=0.0085, Student’s t-test). (d) Animal survival for E0771-bearing mice following the initiation of treatment. Doxorubicin monotherapy improves survival versus the control (*P=0.048, log-rank test), whereas the combination of doxorubicin and losartan enhances this survival increase versus doxorubicin monotherapy (**P=0.014, log-rank test). (e) Animal survival for 4T1-bearing mice following the initiation of treatment. Doxorubicin monotherapy improves survival versus the control (*P=0.045, log-rank test), whereas the combination of doxorubicin and losartan enhances this survival increase versus doxorubicin monotherapy (**P=0.050, log-rank test). (f) Animal survival for AK4.4-bearing mice following the initiation of treatment. The combination of 5-FU and losartan enhances survival versus 5-FU (*P=0.019, log-rank test) or losartan monotherapy (*P=0.027, log-rank test). Animal numbers n=5–6 (E0771, AK4.4 growth), n=3–8 (AK4.4 survival), n=6–7 (4T1). Error bars indicate s.e.m. Statistical tests were corrected for multiple comparisons using the Holm–Bonferroni method.