In this study, we demonstrated the critical role of the FKBPL mechanism in regulating hypoxia- and inflammation-induced endothelial cell function and integrity using an in vivo mouse model of hindlimb ischemia, and an advanced in vitro 3D microfluidic model of human microvascular endothelial cell dysfunction (summarised in Fig. 6). FKBPL has a significant impact on endothelial cell migration, HIF-1α signalling and proteins regulating endothelial cell integrity (e.g. VE-cadherin, CD31, Collagen Alpha-1(XIX), and JCAD). This is the first study to show that FKBPL is a key component of the feedback mechanism that stimulates angiogenesis, in ischaemic tissues of a hindlimb ischemia mouse model, in conjunction with increased HIF-1α and CD31, expression and decreased VE-cadherin, expression.
Using our innovative 3D microfluidics model of endothelial cell dysfunction, we showed that FKBPL suppression by siRNA or hypoxia stimulated HMEC-1 cell migration. Interestingly, under inflammatory conditions, while endothelial cell FKBPL protein expression was increased, cell migration was stimulated, highlighting complex mechanisms that are dependent on the type of stress environment i.e. hypoxia or inflammation. Notably, the FKBPL-based therapeutic peptide, AD-01, can ameliorate the majority of hypoxia- and inflammation-induced changes in endothelial cells by restoring the expression of HIF-1α, VE-cadherin, and CD31, to the control/physiological levels. The most consistent effect was observed for hypoxia-induced endothelial cell impairment via the regulation of HIF-1α. Although AD-01 abrogated the changes in endothelial cell migration induced by hypoxia or low FKBPL expression, this effect was not possible under inflammatory conditions. These findings suggests that under hypoxic conditions, AD-01 may restore the angiogenic balance and maintain the integrity of the endothelial cell barrier, which is essential for proper blood vessel formation, by regulating FKBPL and HIF-1α. Our proteomics analysis identified other anti-inflammatory, anti-angiogenic and vascular remodelling targets of AD-01 under hypoxic conditions, including NF-kß, collagen alpha 1, and junctional cadherin 5 associated proteins, amongst others.
In our in vivo mouse model of hindlimb ischemia, surgical ligation of the femoral artery above the epigastrica and profunda femoris was performed to predominantly promote hypoxia-induced angiogenesis in gastrocnemius muscles 19,20. However, studies have shown inflammation in similar hindlimb ischemia models which is primarily due to ligation below the epigastrica and profunda without removing the femoral artery or ligation proximal to the popliteal artery 43,44. As hypoxia is the primary driving force behind angiogenesis in this model through the regulation of chemokines, including CXCL12 and its receptor, we observed a significant increase in HIF-1α protein expression, which was positively associated with CD31 positive capillaries 19,21,45. Importantly, ischemia-induced angiogenesis resulted in substantial downregulation of FKBPL protein/gene expression, in association with reduced VE-cadherin protein, ICAM, and VCAM gene, expression. Although the suppression of FKBPL appears to be an important mechanism leading to enhanced angiogenesis in ischemia/hypoxia, previous work in fkbpl+/− mice showed that heterozygous knockdown of FKBPL, under normal conditions, stimulated angiogenesis but with impaired vascular integrity and less robust blood vessels 12. Similarly, reduced VE-cadherin expression has been shown to disrupt endothelial cell junctions and reduce vascular integrity 46. In addition, ICAM and VCAM are predominantly stimulated during inflammation to facilitate the recruitment and adhesion of leukocytes to endothelial cells to form and stabilize blood vessels47,48. It remains to be determined whether the reduction in the expression of vascular and inflammatory factors observed in our study is due to surgery, hypoxia, duration of ischemia or reduction in FKBPL expression. Overall, these data from our mouse model of hindlimb ischemia suggest that hypoxia, in conjunction with the upregulation of HIF-1α, may inhibit the expression of FKBPL, resulting in a pro-angiogenic response and a reduction in the integrity and stability of the new vasculature in ischemic tissues.
Similar to in vivo settings, in our 3D microfluidic chips analysis of microvascular endothelial cell dysfunction, FKBPL knockdown or hypoxia induction via DMOG treatment led to a significant increase in HIF-1α and CD31, while reducing VE-cadherin. The FKBPL-based therapeutic peptide mimetic, AD-01, was able to restore the expression of these proteins under hypoxic conditions. HIF-1α plays a critical role in regulating the adaptive response to hypoxia, however, aberrant HIF-1α expression can result in chronic inflammation, oxidative stress, impaired angiogenesis, and energy metabolism, hence contributing to the development of cardiovascular diseases 49–53. Consequently, anti-angiogenic therapeutics including AD-01, which can reverse HIF-1α overexpression and preserve physiological angiogenesis in hypoxia-induced vascular damage are highly desirable.
We also showed that treatment with AD-01 under hypoxic conditions, can abrogate VE-cadherin downregulation and normalise CD31 expression in microvascular endothelial cells. This may result in improved angiogenic responses to ischemia and better perfusion recovery in patients with CVD. CD31 is a transmembrane glycoprotein that is enriched at endothelial cell intercellular junctions and mediates cell-cell adhesion between endothelial cells and adherent leukocytes. It also regulates endothelial cell functions including cell migration, tube formation, and angiogenesis 54–56. Thus, dysregulation of CD31 expression in endothelial cells may result in impaired endothelium and disruption of endothelial junctions, which may exacerbate endothelial dysfunction. In vitro and animal studies have shown that reducing VE-cadherin at the junctions between endothelial cells decreases the integrity of endothelial cells through an increase in endothelial cell permeability, which facilitates the migration of inflammatory cells from blood vessels into tissues 57–59.
In our 3D microfluidic endothelial cell model of inflammation, we used 50% MCM to stimulate the expression of several chemokines and their receptors including CCL2, CCL5, and CX3CL1, which have been shown to regulate inflammation-driven angiogenesis in the vascular system during the development and progression of CVD60–63 .FKBPL has also been shown to regulate inflammation through STAT3, CD44, and NF-κß 18,64,65. Although, previous work demonstrated that AD-01 requires CD44 to inhibit angiogenesis, modulates NF-κß expression in bone marrow derived macrophages in a CD44−/− mouse model, indicating that AD-01 may regulate inflammation through a different pathway in hypoxia 18. While inflammation and hypoxia signalling pathways coexist within the vasculature, the impact of inflammation on HIF-1α remains unclear. Considering that AD-01 regulates HIF-1a expression, we expect that AD-01 might also have anti-inflammatory effects, which merits further investigation.
In contrast to hypoxia, inflammation resulted in an increase in FKBPL expression following 72 h of treatment with MCM. However, AD-01 abrogated the overexpression of FKBPL induced by MCM, but not 2D, and was not able to reverse MCM- stimulated endothelial cell migration in 3D microfluidic chips. A previous study demonstrated that AD-01 is able to negatively regulate FKBPL in cardiomyoblasts after exposure to angiotensin-II for 48 hours 66. These findings suggested that as a FKBPL mimetic, AD-01 has a complex and compensatory mechanism for controlling FKBPL expression in hypoxic and inflammatory environments; some of these mechanisms have been identified through proteomics. Similar to the hypoxia model, MCM treatment also increased HIF- 1α and CD31, while reducing VE-cadherin. However, AD-01 was only capable of restoring the expression of only HIF-1α and CD31 proteins and did not abrogate the decrease in VE-cadherin expression.
In inflammatory conditions, TNF-α stimulates the expression of HIF- 1α via NF-κß-dependent transcription 67,68. Furthermore, we have shown that 24-hour TNF-α exposure also increases FKBPL expression in trophoblasts and endothelial cells 69. FKBPL inhibits NF-κß signalling through its N terminal region, which contains functional and non-functional peptidyl prolyl isomerases 18. There are other members of the FKBP family, including FKBP51 and FKBP52 that interact with p65 complexes to regulate NF-κß 70. Although AD-01 is based on non-functional peptidyl prolyl isomerases of FKBPL terminal, it appears this region of FKBPL may play an important role in regulating HIF-1α expression via NF-κß-dependent pathways. However, additional studies are needed to determine the intricate relationships between FKBPL, HIF-1α, TNF-α, and NF-κß.
In a wide range of CVD cases, hypoxia and inflammation often coexist as features of the tissue microenvironment of the dysfunctional endothelium. We showed that under hypoxia, AD-01 regulates proteins and pathways essential for cellular signalling, tissue remodelling, vascular integrity and inflammation. For example, AD-01 significantly enhances vesicles trafficking, and translocation of nascent secretory and membrane proteins via the activation of the NBAS subunit of NRZ tethering complex and dolichyl-diphosphooligosaccharide-protein glycosyltransferase subunit 2, respectively 71,72. This finding is in line with previous research showing that FKBPL is a chaperone protein and an essential component of the Hsp90/ER complex that contributes to the stability and signalling of the ER 73. AD-01 also enhanced microvascular integrity in conjunction with increasing the expression of the collagen alpha-1(XIX) chain and junctional cadherin associated-5 proteins, under hypoxic conditions. Collagen alpha-1(XIX) is a structural component of the basement membrane that forms and maintains the vascular extracellular matrix and has anti-angiogenic effect by inhibiting cell melanoma cell migration 74,75. Furthermore, AD-01 stabilised the integrity and migratory function of DMOG-treated HMEC-1 cells by likely increasing the expression of JCAD, which is a novel component of endothelial cell-cell junctions that may play a key role in the pathological angiogenic processes 76. Previous research has shown that when JCAD was knocked down, HUVECs exhibit reduced proliferation and migration, increased cell apoptosis, and inhibited Hippo signalling 77. Because AD-01 and FKBPL have been shown to modulate the NF-kß signalling pathway in response to inflammation, we confirmed that AD- 01 might be involved in the regulation of NF-kB signalling via inhibition of the RIG-I-like receptors and VEGF- pathways 18,78–80. Studies have demonstrated that activation of the RIG-I-like receptors stimulates innate immunity and inflammation in endothelial cells by producing type 1 interferon, ICAM, and proinflammatory cytokines, and by inducing the formation of reactive oxygen species, resulting in endothelial dysfunction and vascular disorders 81,82.
Taken together, these findings suggest that FKBPL downregulation during hypoxia and upregulation during inflammation can lead to endothelial cell injury likely through HIF-1α overexpression in vascular systems. AD-01, may be potentially applicable, as a vascular stabilization and anti-inflammatory agent, in a wide range of CVD cases. Using this innovative 3D microfluidic endothelial cell model, we demonstrated that complex pathophysiology can be investigated under different treatments and conditions to mimic human tissue damage in CVD, because the microenvironment is highly controllable and dynamic, and it allows real-time observation of cellular interactions and behaviours.