SRPK1 (serine–arginine protein kinase 1) is a protein kinase that specifically phosphorylates proteins containing serine–arginine-rich domains. Its substrates include a family of SR proteins that are key regulators of mRNA AS (alternative splicing). VEGF (vascular endothelial growth factor), a principal angiogenesis factor contains an alternative 3′ splice site in the terminal exon that defines a family of isoforms with a different amino acid sequence at the C-terminal end, resulting in anti-angiogenic activity in the context of VEGF165-driven neovascularization. It has been shown recently in our laboratories that SRPK1 regulates the choice of this splice site through phosphorylation of the splicing factor SRSF1 (serine/arginine-rich splicing factor 1). The present review summarizes progress that has been made to understand how SRPK1 inhibition may be used to manipulate the balance of pro- and anti-angiogenic VEGF isoforms in animal models in vivo and therefore control abnormal angiogenesis and other pathophysiological processes in multiple disease states.
- age-related macular degeneration
- diabetic nephropathy
- hepatitis C virus
- serine–arginine protein kinase 1 (SRPK1)
- serine/arginine-rich splicing factor 1 (SRSF1)
- vascular endothelial growth factor (VEGF)
SRPK1 (serine–arginine protein kinase 1)
SRPK1 was first described in 1994 following the isolation of proteins that bind phosphorylated SR splicing factors in mitotic HeLa cells extracts . SR proteins contain one or two RNA-binding domains at the N-terminal and a domain with repeated serine–arginine residues at the C-terminal (RS domain) [2,3]. SRPK1 binds and phosphorylates the serine residues in this domain. The first eight serine residues are phosphorylated in a processive manner and the remainder through a distributive mechanism. Phosphorylation is directional, with SRPK1 binding an initiation box (Ser221–Ser225) in the RS domain and then moving towards the N-terminal end .
SRPK1 may be localized both in the cytoplasm and nucleus and translocates between these compartments under various conditions (e.g. cell cycle and stress). It is able to phosphorylate SR proteins both in cytoplasm and nucleus (the significance of which remains incompletely understood) and thus regulate the activity of these proteins by affecting their ability to bind other proteins and/or RNA as well as their intracellular localization . Other members of the SRPK family which are structurally related to SRPK1 and have various substrate specificities and tissue distribution have been described: SRPK2 in mouse  and human  and SRPK3 in mouse  and pig . An alternatively spliced isoform of SRPK1, SRPK1a, has also been reported ; it has a longer N-terminal portion and is so far only known to be expressed in human testis  (for detailed reviews about SRPK1 protein structure and function, see [4,5]).
Over 40 SR proteins have been reported to affect splicing . Given the large number of SR proteins (over 100 putative RS domains have been identified in the genome ) and the possibility that any of these proteins could in certain conditions be phosphorylated by one of the members of the SRPK family, it is possible that SRPKs are involved in regulation of many cellular processes. It is therefore crucial to carry out overexpression and inhibition studies in vivo to fully characterize the essential as well as redundant SRPK1 functions in physiology and pathology.
SRPK1 small-molecule inhibitors
In a quest to find specific inhibitors for SRPK1, Hagiwara and co-workers have screened more than 100000 chemicals using an in vitro kinase assay with an arginine/serine repeats peptide as substrate . They found an isonicotinamide compound they termed SRPIN340, which specifically inhibits SRPK1 (with a Ki<1 μM). It does inhibit SRPK2 as well, but at a much higher concentration and does not inhibit any of the other 140 kinases against which the compound has been tested. The same authors showed inhibition of HIV replication by SRPIN340 . Recently, another group reported inhibition of HCV (hepatitis C virus) replication by SRPIN340 in an in vitro system without affecting proliferation of cells and without cytotoxicity . These studies suggest promise for SRPIN340 use as a potential therapeutic agent.
Control of pro- and anti-angiogenic VEGF (vascular endothelial growth factor) splicing isoforms by SRPK1
VEGF plays major roles in driving angiogenesis and vessel permeability as well as in cell migration and survival . Abnormal levels of VEGF have been implicated in a multitude of diseases, e.g. cancer , several neo-proliferative eye diseases , kidney diseases including diabetic nephropathy , arthritis  and atherosclerosis . It is therefore important to understand the complex regulation of VEGF levels in physiology and disease and try to design novel therapeutic strategies to manipulate quantitatively and qualitatively VEGF levels appropriately.
In 2002, a novel family of VEGF splice isoforms was characterized that result from the use of an alternative 3′ splice site in the terminal exon . VEGF165b, the major isoform of the family, has been shown to act as an antagonist to VEGF165 and a partial agonist of the VEGF receptor in many in vitro and in vivo systems . Recent work has been directed to understand how the choice of the VEGF165/VEGF165b splicing isoforms is regulated. The splicing factor SRSF1 (serine/arginine-rich splicing factor 1) [also known as ASF (alternative splicing factor) or SF2 (splicing factor 2)] has a central role in this regulation by defining the proximal splice site and therefore the expression of the pro-angiogenic VEGF isoforms. Its activity depends on phosphorylation by SRPK1, which is in turn regulated by WT1 (Wilms' tumour-suppressor gene 1) at the transcriptional level . The distal splice site, and hence the expression of the anti-angiogenic VEGF isoforms, is regulated by the splicing factor SRp55 which is modulated through phosphorylation by Clk1 kinase . The choice of proximal or distal splice site is also regulated by extracellular signals such as IGF-I (insulin-like growth factor-I) and TGFβ (transforming growth factor β) . SRPK1 blockade by knockdown or chemical inhibition using SRPIN340 has been shown to switch the balance of VEGF165/VEGF165b splice isoforms in a series of cell lines from both conditionally immortalized primary cells (e.g. podocytes ) and transformed (colon carcinoma ), prostate cancer (A. Mavrou, unpublished work) and melanoma (M. Gammons, unpublished work).
SRPK1 inhibition in tumours
Angiogenesis is essential for tumour growth and several strategies to inhibit angiogenesis have been successfully used in blocking tumour development . Since VEGF is a key driver of angiogenesis, most of these strategies are based on inhibition of VEGF or its receptors, with several small molecules (e.g. Sunitinib) or humanized antibody [e.g. Avastin (bevacizumab)] already approved for clinical use. However, these treatments are far from perfect since they work in only a few types of cancers, only a small number of patients respond to them and there are serious associated side effects (e.g. proteinuria and hypertension in the case of bevacizumab ). Even more worrying are recently published data showing that prolonged exposure to bevacizumab can promote metastasis  or that administration of bevacizumab decreases the uptake of other chemotherapeutic agents in tumours . It is therefore imperative for researchers in the field to develop novel ways of inhibiting angiogenesis during tumour growth and progression. One of these could be to manipulate regulators of VEGF splicing to tilt the balance towards anti-angiogenic isoforms.
As summarized above, SRPK1 is involved in the regulation of the VEGF165/VEGF165b splicing switch through phosphorylation of SRSF1. However, it is possible that several redundant pathways exist that regulate the angiogenic/anti-angiogenic VEGF splicing switch; for instance, in a recent study , it was demonstrated that the transcription factor E2F1 regulates transcription of VEGF as well as the VEGF165/VEGF165b splicing switch through regulation of the splicing factor SC35 (SRSF2). Moreover, SRPK1 is known to have other targets , SRSF1 has been characterized as an oncogene and has been shown to be involved in controlling apoptosis . We have therefore decided to enquire whether knockdown of SRPK1 in vivo inhibits significantly tumour growth, and if so, whether this effect is attributable to alteration of the VEGF165/VEGF165b splicing switch and subsequent inhibition of angiogenesis.
Stable SRPK1 knockdown was achieved in a colon carcinoma cell line (LS174t) and these cells were injected into the flanks of nude mice. SRPK1-knockdown tumours grew significantly more slowly in comparison with their SRPK1-expressing counterpart carcinoma cells. This was because of an inhibition of angiogenesis, as demonstrated by a decrease in microvessel density in SRPK1-knockdown tumours compared with controls . To better understand whether the main reason for the decrease in tumour growth is related to the VEGF165/VEGF165b splicing switch we have overexpressed VEGF cDNA (immune to splicing control) under a CMV (cytomegalovirus) promoter in the SRPK1-knockdown cells and have shown that these SRPK1-knockdown/VEGF-overexpressing cells grow at the same rate as control cells infected with non-targeting lentivirus and transfected with empty vector (S. Oltean, M. Parry and J. Butler, unpublished work).
Recent data from our laboratory suggest that the same effect of SRPK1 inhibition on tumour growth is seen in other types of cancer cell lines, including prostate cancer (A. Mavrou, unpublished work) and melanoma (M. Gammons, unpublished work).
SRPK1 inhibition in models of ocular angiogenesis
VEGF has also been identified as a key regulator of ocular angiogenesis. Exudative AMD (age-related macular degeneration), the most severe form of AMD and visual loss , is characterized by CNV (choroidal neovascularization), the abnormal growth of new vessels from the choroid into the retinal pigmental epithelial cells . Increased VEGF expression, stimulated by hypoxia  and local inflammation  has been identified as the driving force behind CNV progression, leading to loss of photoreceptors, retinal detachment and dense macular scarring [36,37].
In 2006, Ranibizumab (Lucentis), a novel humanized anti-VEGF antibody fragment, was approved by the FDA for the treatment of neovascular AMD, following the demonstration that Lucentis maintained visual acuity in 95%, and improved vision in 40%, of patients with neovascular AMD [38,39]. Since its approval Lucentis has proved successful in clinics and is still regarded as the ‘gold standard’ for treatment of exudative AMD. However, Lucentis is expensive and requires injections directly into the eye every month, an invasive procedure that can lead to increased intraocular pressure  and a risk of severe adverse effects including endophthalmitis . The development of anti-VEGF inhibitors represents a new era in the treatment of neovascular AMD, but novel VEGF inhibitors that avoid the need for intravitreal injections while maintaining potency and specificity, would provide substantial advantages.
We have shown recently that specific low-molecular-mass inhibitors of SRPK1 can successfully suppress VEGF-mediated CNV in vivo. SRPIN340 demonstrated significant dose-dependent inhibition of neovascularization in laser-induced CNV models in both mice  and rats, with beneficial effects seen with high nanomolar concentrations of SRPIN340 (M. Gammons, M. Hagiwara and D.O. Bates, unpublished work). To determine whether the SRPIN340 induced reduction of pathological angiogenesis was accompanied by parallel changes in VEGF splicing in vivo, retinal protein was examined from SRPIN340 and saline-treated eyes following oxygen-induced retinopathy. In SRPIN340-treated eyes a splicing switch was observed favouring production of anti-angiogenic VEGFxxxb isoforms (M. Gammons, M. Hagiwara and D.O. Bates, unpublished work).
Furthermore, SRPIN340 has demonstrated significant CNV suppression when administered daily as a topical eye drop, without any observed adverse effects to the cornea. SRPIN340-mediated inhibition of SRPK1 not only maintains the production of cytoprotective VEGFxxxb isoforms  but was also not toxic to cells even at concentrations as high as 5 mg/ml . This treatment could be an alternative therapy to anti-VEGF-A intraocular injections in patients with exudative AMD, and it may be possible to extend this to diabetic proliferative retinopathy.
Potential for SRPK1 inhibition in diabetic nephropathy
VEGF is also an important vessel permeability factor and essential in maintaining kidney homoeostasis. Numerous studies have now shown that both depletion and overexpression of VEGF within the kidney [34,35] are detrimental and induce various kidney lesions, proteinuria and, in some cases, end-stage renal failure. These studies suggest that VEGF levels have to be very tightly regulated in kidney; indeed, a large number of kidney diseases are associated with either high or low levels of VEGF. These include DN (diabetic nephropathy), with most of the studies reporting association with high levels of VEGF . Similar to the other pathologies described above, there is an acute need for developing novel therapeutic strategies in DN, as no specific treatments are available and the incidence of diabetes and DN continue to rise towards pandemic proportions throughout the world (International Federation for Diabetes, http://www.idf.org).
None of the studies mentioned above have looked at both VEGF165 and VEGF165b isoforms, though it has been shown that they form the total amount of VEGF in human kidneys in approximately equal amounts . Several lines of evidence suggest that GWP (glomerular water permeability) in the kidney is determined by the balance of the two isoforms, e.g. in transgenic mice overexpressing VEGF165b in kidneys GWP is reduced , in double-transgenic mice overexpressing inducible VEGF164 in the kidneys GWP is increased, whereas in triple transgenic mice overexpressing both isoforms GWP is similar to wild-type mice levels (S. Oltean, C.R. Neal, S.J. Harper, D.O. Bates and A.H. Salmon, unpublished work).
In several rodent models, we have recently demonstrated that either transgenic overexpression of VEGF165b or intraperitoneal injection of rhVEGF165b (recombinant human VEGF165b) decreases proteinuria and slows progression of DN as assessed at the histological level . Whether inhibition of SRPK1 has the same effect on DN is unknown, but it is known that SRPK1 inhibition, either pharmacologically or by gene knockdown, can switch splicing of VEGF in podocytes towards the protective VEGF165b isoforms , suggesting this as a potential strategy for inhibition of diabetic nephropathy.
Potential for SRPK1 inhibition in neuropathy
Further evidence also supports VEGF as a major regulator of neuronal function, independent of its actions on the vasculature. In sensory systems, VEGF and VEGF receptor 2 are present in primary sensory neurons in the dorsal root ganglia  and in the spinal cord . In primary sensory neurons, VEGF is a neurotrophic factor, supporting axonal growth and regeneration, has neuroprotective effects [38,40], and is elevated in response to injury of sensory neurons in rodent models of neuropathic pain [39,41]. VEGF has been implicated in nociception, as administration of VEGF165 results in hyperalgesia and allodynia , and abolition of VEGF signalling attenuates neuropathic pain  indicating that elevated VEGF contributes to neuropathic pain. The anti-VEGF and VEGF receptor blockade therapies used in the treatment of many diseases, as discussed above, are reported to result in non-desirable side effects, particularly the development of pain. Pain is consistently reported as an adverse effect with anti-VEGF therapy, occurring in up to 88% of patients (bevacizumab ; telatinib and bevacizumab ). These observations highlight the importance of VEGF signalling in modulating nociception and the need to consider this during the development and use of VEGF-targeted treatment strategies. Both the anti-angiogenic factor VEGF165b and SRSF1 are expressed in primary sensory neurons, and, in contrast with VEGF165, exogenous VEGF165b has anti-nociceptive actions (R. Hulse, unpublished work). Current work is ongoing to investigate the effects of altering the balance between VEGFxxx/VEGFxxxb on nociception, through manipulation of SRSF1 signalling, via inhibition of SRPK1. Inhibition of the SRPK1 pathway would reduce or prevent the increased expression of the pro-nociceptive factor VEGFxxx and increase the proportion of the putative anti-nociceptive VEGFxxxb, to relieve pain. This work will ultimately highlight AS (alternative splicing) and the machinery involved as novel targets for analgesic development.
AS is one of the main processes that influences proteome diversity and it has been estimated in recent studies that up to 94% of multi-exon genes are alternatively spliced in mammals [46,47]. An increasing number of splice isoforms of various genes are reported to be associated and/or drive pathological processes and disease, challenging the scientific community to increase efforts to understand regulation of AS and subsequently try to control abnormal AS for therapeutic benefits. Work summarized in this review shows the potential of targeting VEGF-splicing regulators (i.e. SRPK1) in a quest to obtain a modulation of VEGF isoform expression (Figure 1) which could be more efficient, more cost effective and with fewer side effects than inhibition of the protein itself or its receptors. Future work is needed to address further the biological effects and specificity of SRPK1 inhibition in vivo as well as suitability of SRPK1 inhibitors for therapeutic use.
A.H.S. was supported by a Medical Research Council and Academy of Medical Sciences Clinician Scientist Fellowship [grant number G0802829]. R.H. was supported by a project grant from Diabetes UK [grant number 11/0004192]. S.O. was supported by a project grant from British Heart Foundation [grant number PG 11/20/28792].
RNA UK 2012: An Independent Meeting held at The Burnside Hotel, Bowness-on-Windermere, Cumbria, U.K., 20–22 January 2012. Organized and Edited by Raymond O'Keefe and Mark Ashe (Manchester, U.K.).
Abbreviations: AMD, age-related macular degeneration; AS, alternative splicing; CNV, choroidal neovascularization; DN, diabetic nephropathy; GWP, glomerular water permeability; HCV, hepatitis C virus; SRPK1, serine–arginine protein kinase 1; SRSF1, serine/arginine-rich splicing factor 1; VEGF, vascular endothelial growth factor
- © The Authors Journal compilation © 2012 Biochemical Society