Differences in proliferation rate between CADASIL and control vascular smooth muscle cells are related to increased TGFβ expression

Abstract Cerebral autosomal‐dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is a familial fatal progressive degenerative disorder. One of the pathological hallmarks of CADASIL is a dramatic reduction of vascular smooth muscle cells (VSMCs) in cerebral arteries. Using VSMCs from the vasculature of the human umbilical cord, placenta and cerebrum of CADASIL patients, we found that CADASIL VSMCs had a lower proliferation rate compared to control VSMCs. Exposure of control VSMCs and endothelial cells (ECs) to media derived from CADASIL VSMCs lowered the proliferation rate of all cells examined. By quantitative RT‐PCR analysis, we observed increased Transforming growth factor‐β (TGFβ) gene expression in CADASIL VSMCs. Adding TGFβ‐neutralizing antibody restored the proliferation rate of CADASIL VSMCs. We assessed proliferation differences in the presence or absence of TGFβ‐neutralizing antibody in ECs co‐cultured with VSMCs. ECs co‐cultured with CADASIL VSMCs exhibited a lower proliferation rate than those co‐cultured with control VSMCs, and neutralization of TGFβ normalized the proliferation rate of ECs co‐cultured with CADASIL VSMCs. We suggest that increased TGFβ expression in CADASIL VSMCs is involved in the reduced VSMC proliferation in CADASIL and may play a role in situ in altered proliferation of neighbouring cells in the vasculature.


| INTRODUCTION
Cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is the most common form of hereditary small vessel disease and is the most common genetic cause of stroke and vascular dementia in adults. 1,2 The clinical presentation of CADASIL is characterized by migraine with aura, transient neurological symptoms, mood disturbances and cognitive impairment. [3][4][5][6][7] Pathological characteristics evident in the brain of patients with CADASIL include deposits of granular osmiophilic material (GOM) in close vicinity to the basement membrane that surrounds vascular smooth muscle cells (VSMCs) of parenchymal arterioles, 8,9 and degeneration of VSMCs 10,11 in small and medium-sized arteries. 12,13 Subsequent to VSMC degeneration, fibrotic thickening of the vessel walls and narrowing of the lumina of small penetrating arterioles occur predominantly in cerebral white matter. 14 The resulting blood flow restriction leads to characteristic ischaemic changes, white matter loss and lacunar infarcts. 14 White matter changes observed via magnetic resonance imaging (MRI) are a characteristic feature of CADASIL that have been utilized as an indication of the disease. 15,16 CADASIL is caused by mutations in the NOTCH3 gene. NOTCH3 is a type I transmembrane receptor belonging to the Notch signalling family, one of a group of "elite" intracellular signalling pathways. 17,18 Currently, more than 230 different CADASIL-causing mutations in the NOTCH3 gene have been detected, 19 which are either missense mutations or small in-frame deletions in the epidermal growth factor repeats of NOTCH3. These mutations affect the number of cysteine residues, leading to an unpaired cysteine residue. [20][21][22][23][24] In CADASIL, pathological observations have shown abnormality in numerous cells within neurovascular units, including VSMCs and endothelial cells (ECs). 25,26 Therefore, alterations to each cell type within the neurovascular unit can affect the vulnerability of blood vessels in CADASIL. 25 In CADASIL, the outer layer of the vessels consisting of VSMCs becomes thinner, while the endothelial layer that is surrounded by the VSMCs becomes swollen and loses its integrity due to gap junction loss. 25,26 This leads to microbleeds, occlusion and/or thrombosis. 27,28 It has been proposed that the degeneration of the smooth muscle layer in arteries/arterioles of patients with CADASIL is the result of increased VSMC death. 11,29 There is meagre evidence that the thinning of the tunica media is due to VSMC loss or degeneration by cell death. In a recent study, we suggested that the reduced number of cells is rather due to impaired VSMC proliferation. 29 We showed that VSMCs from CADASIL patients with the R133C mutation in NOTCH3 have a lower proliferation rate than control VSMCs. 29 Transforming growth factor-b (TGFb) is a cytokine that has multicellular function in the cells and has a great involvement in the fibrosis of several diseases. The role of TGFb in cell proliferation is well-established. 30 Dysregulation of the TGFb signalling pathway has been proposed in CADASIL by showing increased levels of the TGFb pro-domain. 31 Furthermore, latent TGFb binding protein 1 (LTBP-1) has been found co-localized with NOTCH3 extracellular domain deposits in CADASIL tissues. Transforming growth factor-b signalling has also been demonstrated to play a role in the promotion and maintenance of the contractile phenotype of VSMCs. Three TGFb isoforms (TGFb1, TGFb2 and TGFb3) have been identified in mammals, which are encoded by different genes but have high structural homology. Transforming growth factor-b binds to membrane receptors bearing serine/threonine kinase activity, namely TGFb receptors (TGFbR). The TGFb isoforms have similar cellular signalling targets on the TGFb receptors which exist in I, II, III and V subtypes. 32 The expression level of TGFb-receptors is important for the TGFb signalling pathway. 33 This study aimed to investigate why VSMC proliferation is impaired in CADASIL. Using immortalized human umbilical, placental and cerebral VSMCs derived from individuals carrying the R133C mutation or control NOTCH3, we report a role of TGFb in VSMC and EC proliferation rates, suggesting it is a critical mechanism involved in CADASIL pathophysiology.

| Cell lines and genotyping
Patient-derived VSMC cell lines: umbilical cord (Umb)-and placental (Pla) VSMCs were established from blood vessels of genetically verified patients with CADASIL and control subjects as previously described. 34,35 The genotyping of the cerebral arterial (Cer) VSMC cell line 35 was also verified ( Figure S1A).

| Cell culture
Patient-derived VSMC cell lines: Umb-, 29 Pla-and CerVSMC and human foreskin fibroblasts were cultured as described previously. 35

| Cell proliferation rate examination
CellTrace carboxyfluorescein succinimidyl ester (CFSE) (Invitrogen) was used to enable the measurement of VSMC proliferation. Carboxyfluorescein succinimidyl ester is a membrane permeant fluorescein-based dye that can be used to track cell division due to the progressive halving of the fluorescence intensity of the dye in cells after each division. Briefly, CADASIL and control VSMC were grown to 90% confluence and harvested after washing twice with PBS. The cells were centrifuged and re-suspended with 1 mL pre-warmed PBS followed by labelling with 1 lmol/L CFSE for 10 minutes at 37°C. PANAHI ET AL.

| 3017
Equal numbers of CFSE-labelled cells were incubated for either 3 or 7 days. Unlabelled cells were used as negative controls. Flow cytometry was performed using a FACSCalibur ™ Cytometry, and data were analysed using the CellQuest software (BD Biosciences). The proliferation rate of VSMCs was also investigated using immunofluorescent staining of the proliferation-associated Ki67 protein as previously described. 29 The quantification of Ki67 positive cells was based on the evaluation of at least 500 cells.

| Exposure of fibroblasts and VSMC to CADASIL PlaVSMC-conditioned medium
Human fibroblasts and UASMC were exposed to media that CADA-SIL PlaVSMC cells were grown in conditioned medium for 24 hour, and 3 and 7 days. The PlaVSMC-conditioned medium was applied either undiluted or diluted 1:1 with complete medium. The proliferation rate of VSMCs was investigated using immunofluorescence staining of the proliferation-associated Ki67 protein as described previously. 29 The ratio of Ki67 positive cells was based on the evaluation of at least 500 cells. Experiments were performed in triplicate.

| Transforming growth factor-beta (TGFb) antibody
Transforming growth factor-b pan-specific neutralizing antibody was derived from recombinant human TGFb1, porcine platelet-derived TGFb1.2, porcine platelet-derived TGFb2 and recombinant amphibian TGFb5 porcine platelet-derived TGFb1 and 2, which detects TGFb1,TGFb1.2, TGFb2, TGFb3 and TGFb5 (AB-100-NA, R&D systems). Vascular smooth muscle cells was seeded at 5 9 10 4 cells/mL in 24 well plates and cultured overnight. The medium was then changed and cells were incubated in the presence or absence of 12 lg/mL TGFb-neutralizing antibody for 24 hour. Vascular smooth muscle cells or ECs were fixed and stained with Ki67 antibody as described earlier.

| Co-culture of VSMCs and ECs in transwell system
To examine the effect of TGFb produced by VSMC on EC proliferation rates, Pla-and CerVSMC were co-cultured with ECs in a noncontacting co-culture transwell system (Pore size 0.4 lm, Life Technologies). Endothelial cells were plated at 5 9 10 4 cells/mL in 6-or 24-well plates. Pla-and CerVSMCs were seeded at 5 9 10 4 cells/ mL onto the membrane of transwell cell culture inserts and allowed to grow overnight. After 24 hour, the transwell insert membrane containing the VSMCs was placed into plates containing ECs. Vascular smooth muscle cells were treated with a TGFb-neutralizing antibody (12 lg/mL).    (Table S1, data not shown, respectively).

| Statistical analysis
We found an increase expression of all TGFb isoform families

| TGFb secreted by CADASIL VSMC influences proliferation of endothelial cells
To investigate whether TGFb produced by CADASIL VSMC affects EC proliferation, we co-cultured ECs with CADASIL or control Plaand CerVSMCs using the transwell system ( Figure 4). Endothelial cells co-cultured with CADASIL Pla-and CerVSMCs exhibited a lower proliferation rate than those co-cultured with control VSMCs ( Figure 4C and D). Neutralization of TGFb rescued the low proliferation rate of ECs co-cultured with CADASIL Pla-and CerVSMCs (Figure 4A-D, P < .05). The proliferation rate of ECs did not change in the presence of TGFb-neutralizing antibody as shown by confocal microscopy ( Figure S4A) and neither did their morphology (Figure S4B).
To examine possible reasons for decreased EC proliferation, we examined the expression of TGFb receptors in ECs co-cultured with CADASIL VSMCs. Our qRT-PCR data showed that TGFbR3 was down-regulated in ECs co-cultured with CADASIL VSMCs compared to those cultured with control VSMCs ( Figure S5). TGFb and TGFb receptor 1 and 2 were unaltered by the presence of CADASIL VSMCs.

| DISCUSSION
One of the major pathological findings of CADASIL is degeneration of small arterioles due to the loss of myocyte cell coverage. 11 Previously, we have demonstrated that VSMCs derived from patient with CADASIL have a lower proliferation. 29 In this study, we have revealed that the lower proliferation rate of these cells affects neighbouring ECs. In addition, we demonstrated that this inhibition of proliferation occurs in a paracrine fashion by inhibiting cell to cell contact using a transwell system. This paracrine mechanism seems to be due to a compound that is secreted by VSMC, and one cytokine that could potentially explain this phenomenon in CADASIL is TGFb. By blocking this cytokine using an anti-TGFb antibody, we were able to restore CADASIL VSMCs and ECs co-cultured with CADASIL VSMCs to a proliferation rate that resembles the proliferation rate of ECs co-cultured with control VSMCs.
Transforming growth factor-b has been shown to be involved in differentiation/proliferation of VSMCs. 30,37 Dysregulation of the TGFb signalling pathway has been proposed to play a role in CADA-SIL as observed by an increased levels of the TGFb. 31 The TGFb family implication in fibrosis is most interesting as fibrotic thickening of small vessels in CADASIL is widely reported. 38 It has been suggested that the damage of VSMCs in CADASIL induces secondary fibrosis with consequent thickening of the walls and narrowing of the lumen of cerebral arteries. 19 In this study, we found elevated expression level of TGFb isoforms (TGFb1, TGFb2 and TGFb3) in Plaand CerVSMCs, where the increased level of TGFb3 was more obvious as compared to the TGFb1 or TGFb2. TGFb isoforms are expressed in the central nervous system and among them, TGFb1 has been widely considered as an injury-related cytokine which is a F I G U R E 2 Proliferation rate of foreskin fibroblasts cultured in Cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) PlaVSMC-conditioned medium. A, Human fibroblasts were incubated with PlaVSMC-conditioned medium for 1, 3 and 7 days, stained with Ki67 antibody and analysed by confocal microscopy. B, Umbilical artery smooth muscle cells (UASMCs) were also incubated with CADASIL and control PlaVSMCsconditioned medium. C, D, qRT-PCR analysis of TGFb3 gene in Pla-and CerVSMCs. The expression of TGFb3 gene was normalized to the endogenous control gene; HPRT1, and the RQ (Relative Quantitation) was calculated using control VSMC normalized to 1. **P < .01, ***P < .001. The results are representative of three independent biological replicates (n = 3). One-way ANOVA followed by Bonferroni's post hoc test was used for statistical analysis crucial regulator of nervous system physiology and its vasculature. TGFb3. [40][41][42] TGFb3 has also been detected in uninjured epithelium.
However, distinct effects of vascular-derived TGFb3 have not been elucidated in small vessel diseases and CADASIL. In the current study, we found an up-regulation of TGFb3 gene expression in CADASIL VSMCs. Other TGFb-isoforms genes were also expressed higher in CADASIL VSMCs compared to control VSMCs. Variability in expression of the TGFb3 gene could result in enhanced myofibroblast activity, proliferation and induction of ECM synthesis. 43,44 A previous study has identified elevated genes related to ECM proteins and collagens in CADASIL. 45 There is a possibility that the observed increased secretion of TGFb-isoforms, in particular TGFb3 in our study, could be involved in induction of ECM protein synthesis and fibrosis formation that disrupts the normal architecture of vasculature in CADASIL. Whether fibrosis is a general response to degeneration of VSMCs or because of the activation of TGFb-specific fibrotic pathways is still to be examined.
Degenerative changes in capillary vessels are also involving ECs in CADASIL and showing endothelial abnormalities which suggests a secondary character of endothelium damage in CADASIL. 46 EC function in blood vessels is regulated by coordination of vascular VEGF, Notch and TGFb. 47 Studies have shown that signalling of TGFb/ ALK1-Smad1/5 stimulates EC migration and proliferation. 48 Inhibition of endoglin in ECs, which is a part of the TGFb receptor complex, potentiates TGFb/ALK1 signalling, 49,50 resulting in reduced proliferation. 51,52 TGFb binds to its specific receptors including TGFbR3. TGFbR3 is one of the proteins that have a more indirect role in TGFb signal transduction. 33 The down-regulation of TGFbR3 on EC potentiates the inhibitory effect of TGFb on EC migration and growth. 50 We investigated whether TGFb receptor expression may have an impact on EC proliferation in ECs co-cultured with CADASIL VSMCs. Our qRT-PCR data showed down-regulation of TGFbR3 in ECs co-cultured with CADASIL VSMCs (Table S2, Figure S5  We believe that increased TGFb3 reflects an inflammatory condition and even an involvement of TGFb in fibrosis in CADASIL. Support for our hypothesis is the outcome of several studies that showed NOTCH3 and TGFb1 signalling play a key role in the pathogenesis and progression of chronic cardiovascular disease. 53 NOTCH3 was shown to be an important protective factor against cardiac fibrosis in a myocardial infarction model, and the protective effect of NOTCH3 is attributable to its action on TGFb1/Smad3 signalling. This might also occur in patients with CADASIL, where the signalling of TGFb pathway and the effect of gene expression switch the normal function of its action. The role of activated NOTCH signalling and TGFb has also been shown in liver fibrosis. 54 Once the intracellular domain of the NOTCH protein (NICD) is transported into nucleus, it can interact with a number of different transcription factors, including FoxH1, c-Jun, c-Fos, Gli-3 and others and control expression of a large number of target genes, including, for example, genes involved in cell cycle control, extracellular matrix regulation and mesoderm specification. The TGFb signalling pathway is known to be involved in modulation of cyclin genes. 55 We have also observed down-regulation of cyclins gene expression in CADASIL VSMCs, suggesting that TGFb signalling pathway may exert its effect on cyclin genes.
In CADASIL, the cells in the vasculature can be exposed to complex changes, such as aggregated NOTCH3, and GOM. In this study, we have implicated TGFb and TGFb receptors as additional players in CADASIL disease. Therefore, more detailed studies are needed to elucidate the interplay of TGFb signalling in the vasculature in CADASIL. In summary, our data provide evidence for direct involvement of TGFb in proliferation of VSMCs and ECs. We believe that our findings provide a better understanding of the pathogenesis of CADASIL disease.

ACKNOWLEDG EMENT
The human foreskin fibroblast was a kind gift from Professor Outi Hovatta, (CLINTEC) at Karolinska Institutet. This research was supported by grants from Gun and Bertil Stohne's Foundation, Olle Engkvist Byggm€ astare Foundation and Foundation for Gamla Tj€ anarinnor.
F I G U R E 4 Transforming growth factorb secreted by Cerebral autosomaldominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) VSMC influences proliferation of endothelial cells (ECs). A, Representative confocal images of ECs co-cultured with PlaVSMCs and B, CerVSMCs in the presence (anti-TGFb+) (lower panels) or absence of TGFb-neutralizing antibody (anti-TGFb-) (upper panels), stained with Ki67 (green). C, D, Quantitative analysis of Ki67 positive cells prior to and after TGFbneutralizing antibody.*P < .05. Scale bar = 10 lm. The results are representative of three independent biological replicates (n = 3). Student t-test was used for twogroup comparisons