All LDLR vectors contained identical expression cassettes of rabbit LDLR with the synthetic liver-specific a1-microglobulin/bikunin (ABP) / thyroid hormone-binding globulin (TBG) enhancer/promoter (LSP). Green fluorescent protein (GFP) (I-III) and β-galactosidase (LacZ) (II) were used as control genes (Figure 4).
Figure 4. Vector constructs used in the studies.
LV vectors were produced by co-transfection of vector plasmid and helper plasmids into 293T-cells using FuGENEÒ transfection reagent or calcium-phosphate precipitation. Fresh media was changed 16-20h after transfection, and media was collected 24, 48 and 72 h after the first media change. Collected media was filtered through 0.22µm filters and concentrated by two rounds of ultracentrifugation. The resulting virus pellet was resuspended in Hanks’ buffer and stored at -80°C. The titer of the virus was determined by HIV-1 p24 ELISA assay (Alliance, NEN Life Science Products, USA).
AAV2 vectors were produced as previously described (Zolotukhin et al. 1999).
Subconfluent 293T-cells were transiently transfected with calcium-phosphate and appropriate vector and helper plasmids (Table 2). Cells were harvested 48-72h after
transfection. Viral vector was released from cells by 3 freeze-thaw cycles. The vector-containing media was then purified by iodixanol-gradient centrifugation and heparin-affinity chromatography. Fractions containing the purified vector were collected and dialysed against PBS. The purified vector was stored in PBS at -70°C until use.
AAV9 vectors were prepared at the Vector Development Core Lab at UCSD as previously described (Huusko et al. 2012). AAV9 helper plasmid was obtained from Dr. Wilson (U. Penn). Cell lysates prepared at 72h after transfection were treated with benzonase and viruses were pelleted through 25% sucrose-cushion ultracentrifugation. The pellets were resuspended and the viruses were further purified through anion-exchange column chromatography (Q-Sepharose, GE Health Science) followed by concentration by ultracentrifugation through a 25% sucrose-cushion. The final pellets were resuspended in 10mM Tris-HCl, pH 7.9, 1mM MgCl2, 3% sucrose. Virus titers for both serotypes were determined by measuring the genome copies by Real-time quantative PCR (qPCR)
Table 2. Plasmids used in the production of viral vectors.
Plasmid virus Description study
psub-CMV-EGFP-WPRE AAV2 EGFP vector plasmid III
pDG AAV2 Ad/AAV helper plasmid II,III
pAd helper AAV9 Ad helper plasmid II
pRep2-Cap9 AAV9 AAV9 cap plasmid II
ABP/TBG-rLDLR AAV2/AAV9 LDLR vector plasmid I,II
psub-CMV-sMSR-WPRE AAV2 sMSR vector plasmid III
pRSV-rev LV HIV-1 rev packaging plasmid I,II
pMD.G LV VSV-G envelope plasmid I,II
pMDLg/pRRE LV HIV-1 gag-pol packaging plasmid I,II
LSP-LDLR LV LDLR vector plasmid I,II
LSP-GFP LV GFP vector plasmid I,II
4.2 CELL CULTURE EXPERIMENTS 4.2.1 Degradation of I125-LDL, -oxLDL or -acLDL
HepG2 cells or Raw 264.7 macrophages were transduced with LV-LDLR (MOI 5; I), AAV2-rLDLR (MOI 10 000; II) or cells were preincubated with conditioned media from AAV2-sMSR transduced cells (III). Cell were pre-incubated in 20% FBS for 24h to shut down endogenous LDLR activity. Degradation of I125-LDL (I, II), I125-oxLDL
of the iodinated LDLs. Specific degradation of the I125-LDL was measured as previously described (Ylä-Herttuala et al. 1989).
4.2.2 Binding and uptake of fluorescent DiI-LDL
HepG2 hepatoma cells or WHHL rabbit fibroblasts were transduced with LV-LDLR or LV-GFP (MOI 5; I) in six-well plates and cultured for 1-4 weeks. Cells were preincubated in 20% FBS for 24h to shut down endogenous LDLR activity. Cell culture media with 10% LDPS and 10 µg/ml DiI-LDL was added to the cells and they were incubated for 4-12h. Cells were fixed with 1% PFA. LDL uptake was visualized using fluorescence microscopy and quantified by flow cytometry.
4.3 ANIMAL EXPERIMENTS 4.3.1 In vivo gene transfers
Vectors were injected directly into the portal vein through a 22G catheter. Rabbits were anesthetized with phentalnylphluonizone (Hypnorm 0.2 ml/kg, Janssen Pharmaceutica, Belgium) and midazolam (Dormicum 2ml/kg, Roche) (I and III) or with ketamine (Ketalar 20 mg/kg, Pfizer) and medetomidine (Domitor 0.3 mg/kg, Orion, Finland) (II). Laparotomy was performed, and the portal vein visualized.
Vectors were injected via a 24G cannula into the portal circulation. After removal of the cannula the puncture site was observed to make sure no bleeding occurred after which the laparotomy wound was closed in two layers. Carprofen (Rimadyl 4 mg/kg, Pfizer) was given as a postoperative analgesic. Animals were sacrificed 1 month (I, II), 6 months (III), 1 year or 2 years (I,II) after gene transfer. Study groups and doses used are listed in Table 3. All animal experiments were approved by the Institutional Animal Care and Use committee of the University of Kuopio or the Animal Experimental Board.
Table 3. Vectors, doses and time-points used in the study.
vector dose time-point
LV-LSP-LDLR 1x109 IU (10µg gag antigen) 1/2 year
CMV-GFP 1x109 IU (10µg gag antigen) 1/2 year
LSP-GFP 1x109 IU (10µg gag antigen) 1/2 year
AAV2-LDLR 1x1012 vg 1 month / 1 year
AAV2-control 1x1011 vg 1 month / 1 year
AAV9-LDLR 1x1011 vg 1 month / 1 year
AAV9-control 1x1011 vg 1 month / 1 year
AAV2-sMSR 1x1011 vg 6 months
AAV2-EGFP 1x1011 vg 6 months
4.3.2 Uptake of DiI-LDL in liver
For quantification of the uptake of 1, 1'-dioctadecyl- 3,3,3',3'-tetramethylindocarbocyanine perchlorate (DiI; Invitrogen) labelled LDL in the liver, 5µm thick sections were cut with a cryostat from liver samples freshly frozen in OCT compound. Fluorescent images were taken using an Olympus U-RLF-T burner. Five images were taken at 100x magnification from each section. The area of fluorescence was quantified using analySIS software (Soft Imaging System) from 5 sections from each animal.
4.4 LIPIDS AND LIPOPROTEIN METABOLISM 4.4.1 Analysis of blood samples
Blood samples were collected from fasted animals weekly for the first month after gene transfer, and monthly thereafter. Alanine aminotransferase (ALT), alkaline phosphatase (APT), aspartate aminotransferase (AST), C-reactive protein (CRP), bilirubin (Bil), total cholesterol (TC), low-density lipoprotein (LDL), high-density lipoprotein (HDL), albumin (Alb) and triglycerides (TGs) were analyzed at Kuopio University Hospital or MoVet veterinary service laboratory (Kuopio, Finland).
4.4.2 Isolation and DiI-labeling of LDL
LDL was isolated from fresh human plasma, obtained from the Finnish Red Cross, by sequential density gradient ultracentrifugation as previously described (Yla-Herttuala et al. 1989). Isolated LDL was labelled with DiI at +37°C for 15h, the density was raised to 1.1 g/cc with dry KBr and DiI-LDL was isolated by ultracentrifugation in a SW40 rotor at 37000 rpm for 20h at 4°C. Labelled LDL was extensively dialyzed against NaCl/0.01% EDTA buffer and filtered before use.
4.4.3 En face -lesion area
After sacrifice rabbits were perfused via the left ventricle with 1xPBS to clear the aorta of blood. The aorta, from the iliac bifurcation to the first intercostals arteries, was dissected out, further cleaned under a dissection microscope, cut open and pinned down on a black plastic mat. The aortas were stained with Oil Red –O (III) or Sudan IV (II) which stain neutral lipids. Stained aortas were photographed and the percentage of aortic area stained red was quantified using analySIS software (Soft Imaging System).
4.5 TISSUE ANALYSES
sections were analyzed in a blinded fashion by a hepatopathologist (I, II) to assess possible pathologies in the livers after gene transfers.
Table 4. Antibodies used in the studies.
Antibody Clone Dilution/concentration Source Original
publication
β-galactosidase - 1:200 Promega II
rabbit
macrophages RAM11 1:1000 Dako I-III
Anti-flag tag M2 10µg/ml Sigma III
pan-cytokeratin AE1/AE3 1:200 Santa Cruz
Biotechnology 4.6.1 Extraction of nucleic acids
Total RNA was extracted from snap-frozen tissues using TRI-reagent according to manufacturers instructions . DNA was extracted either from snap-frozen tissue with proteinase K digestion followed by phenol-chloroform extraction (I-III), or from FFPE-tissue with the same method after removal of paraffin with xylene and hydration in a graded alcohol series. After extraction, DNA was suspended in H2O and RNA in H2O with RNAse inhibitor. Concentration and purity of nucleic acids were measured with NanoDrop.
4.6.2 PCR and RT-PCR (I)
PCR was carried out on 2-3 µg of gDNA. For RT-PCR 0.5 µg (cell samples) or 2 µg (tissue samples) of DNase treated total RNA was used for cDNA synthesis and further amplified using the Titan One Tube RT-PCR system (Roche Diagnostics GmbH, Germany) with 20 pmol of forward and reverse primers for the first PCR reaction and 15 pmol of primers for the nested PCR. Primers are listed in Table 5.
4.6.3 qPCR and RT-qPCR (II, III)
qPCR for copy number analysis (II, III) was carried out on 50 or 500 ng of genomic DNA. Results are expressed as genome copies/µg DNA. Primers and probe assays used in the studies are listed in table 5.
For RT-qPCR a total of 1 µg RNA was used for cDNA synthesis after DNAse treatment using random hexamer primers and RevertAid RT enzyme. The relative expression levels of Cyr61 and rLDLR in liver samples from the different treatment groups were measured according to the manufacturers´ protocol. The results were normalized to rabbit β-actin.
Table 5. Primers and probe assays used in the studies.
Target fwd primer rev primer probe
LSP GCCTCTGCTTTTGTACAACTTTCC AGTTCTCACTATTGGGCCAAACAG AAAACTGCCAATCCC WPRE ATACGCTGCTTTAATGCCTTTG GGGCCACAACTCCTCATAAA TCATGCTATTGCTTCCCGTATGG
CT
rLDLR GTATCTCCTACAAGTGGGTGTG GACTTGCAGGTGAGA GACAT /56-FAM/CG GCT CGG A/Zen/C GAG TGG GAG CAG /3IABkFQ/
rActB GGACCTGACCGACTACCT GTAGCACAGCTTCTCCTTGAT
/56-FAM/ATGAAGATC/Zen/CTCACGG AGCGCG /3IABkFQ/
sMSR TACAAGGACGACGATGAC CCAGTGGGACCTCGATCTCC -
sMSR
The expression of sMSR protein from the liver was analyzed from serum samples with enzyme linked immunosorbent assay as described previously (Jalkanen, Leppänen, Pajusola, et al. 2003)
4.7 STATISTICAL ANALYSES
Results are shown as means ± SD or SEM as indicated. Statistical significance was evaluated using t-test, one-way ANOVA with Tukey’s multiple comparison test, two-way ANOVA with Bonferroni post-test for multiple comparisons or mixed models analysis where appropriate. Statistically significant was assigned at p<0.05.
Statistical analyses were performed with GraphPad Prism (I-III) or R statistical
5 RESULTS
5.1 INTRAPORTAL GENE TRANSFER OF LV-LDLR LEADS TO A LONG-LASTING REDUCTION IN SERUM TOTAL
CHOLESTEROL IN WHHL RABBITS (I) 5.1.1 In vitro analysis of vectors
LVs expressing the rabbit LDLR or GFP as a control were produced to high titers and used for in vitro transduction of HepG2 cells and WHHL fibroblasts to analyze the expression and functionality of the vectors, and specificity of the liver-specific synthetic ABP/TBG promoter (I). LVs with the liver-specific promoter efficiently transduced HepG2 cells but not WHHL fibroblasts, and showed increased degradation of I125-labelled LDL showing functionality of the transduced LDLR.
5.1.2 In vivo experiments
Intraportal gene transfer of 1x109 TU of LV-rLDLR led to sustained expression of rLDLR for over two years (I). Expression was limited to the liver while provirus was detected in non-target tissues such as spleen and lungs without evidence of gene expression (I). Total cholesterol levels in LV-LDLR transduced animals showed a steady decrease from 4 weeks after gene transfer with total cholesterol levels 34±10%
lower compared to GFP-transduced control animals two years after gene transfer.
Gene transfers were well tolerated. A transient elevation in liver enzymes (AST and ALT) was seen 2-4 weeks after gene transfer related to the liver biopsies collected at these time point. Levels were normalized within a week after the biopsy and remained within the normal range for the duration of the study. Normal CRP levels suggested a lack of infection or inflammation which was further confimed by the lack of increases in RAM-11 positive macrophages or T-cells in liver biopsy samples taken 2-4 weeks after gene transfer (I).
5.2 LV-rLDLR IS MORE EFFICIENT THAN AAV2- OR AAV9-rLDLR IN THE TREATMENT OF FH (II)
5.2.1 Expression of transgenes and clinical chemistry
In order to find the optimal long-term vector for liver gene therapy of FH, AAV-and LV vectors were compared in WHHL rabbit liver for efficacy and safety. AAV vectors encoded the same liver-specific ABP/TBG promoter and rLDLR that was previously used in the LV-vectors. Both AAV serotype 2 and 9 vectors were produced and tested for expression and functionality in vitro (II).
Intraportal gene transfers of AAV-and LV vectors were carried out without complications and rabbits were followed for one year after gene transfer. RT-qPCR from liver samples collected a year after gene transfer showed expression of rLDLR in all groups (II). Analysis of non-target tissues for biodistribution of vectors showed that vector was spread after intraportal gene delivery to spleen, heart, lung, kidney and gonads. The widest distribution was seen after LV-rLDLR gene transfer with generally higher copy numbers (CNs) than after AAV2 or AAV9 gene transfers (II).
Total cholesterol, LDL, HDL and TGs were monitored from fasting serum samples regularly after intraportal gene transfers. In the first five months after gene transfer AAV9-rLDLR and LV-rLDLR were equally efficient at lowering total cholesterol levels (II). However, starting 20 weeks after gene transfer a decline was seen in the AAV9 control group, and TC levels were 38 ± 13% below the pre-treatment values after one-year follow-up compared to a decrease of 37 ± 4% in the AAV9-LDLR group (p=0.062). In LV-rLDLR group total cholesterol was 47 ± 9%
below pre-treatment values one year after gene transfer compared to an increase of 6 ± 15% in LV-control animals (p=0.033). AAV2-rLDLR did not lead to a reduction in total cholesterol values but a significant increase compared to AAV2 control groups was seen (II).
Analysis of vector genome copy numbers (GCs) in liver 1 year after gene transfer of AAV-rLDLR or LV-rLDLR gene transfers revealed the greatest GCs in AAV2-rLDLR group, followed by AAV9-LDLR and LV-LDLR (II). Unexpectedly, the GCs in the biodistribution samples after LV-LDLR gene transfer had higher copies than the liver. The greatest number of copies was seen in the spleen, followed by the heart and lungs. No copies were detected in the kidney, and very low copies in the gonads.
5.2.2 Histology and characterization of bile-duct proliferation
Hematoxylin and eosin (HE)–stainings of liver samples from all groups were analyzed by a liver pathologist and changes were scored according to severity and/or prominance. Generally, all samples showed varying levels of micro- and macrovesicular steatosis and bile-duct intraepithelial lymphocytes (Figure 5).
Ballooning degeneration was seen in all AAVgroups one month after GT but was less prominent at the end of the follow-up. The most severe ballooning degeneration was seen in the LV-control group at the end of the follow-up (II)
An unexpected finding of bile-duct proliferation was seen in the AAV2-LDLR group one year after GT. The finding was regarded as a reactive change according to the pathologist. No malignancies were detected. Further immunohistochemical characterization of the aberrant tubules revealed that they were positive for pan-CK, a marker of epithelial cells. A proliferation marker, Ki-67, revealed some proliferating cells in the area with the bile-duct proliferation but the frequency of Ki-67+ cells did
Figure 5. Liver histology and Ki-67 and pan-cytokeratin staining one year after intraportal gene transfers of AAV2-, AAV9-, and LV-rLDLR and control vectors. Scale bar 100µm.
5.2.3 Atherosclerosis after AAV2- and AAV9 gene transfers (unpublished findings)
Atherosclerosis was also evaluated after AAV-rLDLR gene transfers since it is the major consequence of persistently elevated serum cholesterol levels in FH. The extent of atherosclerosis in aortas of AAV2-and AAV9-transduced rabbits was analyzed from Sudan IV-stained, longitudinally opened aortas. No differences in en face lesion area were seen between the different AAV-groups (Figure 6).
Figure 6. Atherosclerotic lesion area one year after AAV-mediated gene transfers of rLDLR or LacZ.
5.2.4 Uptake of DiI-LDL in liver (unpublished findings)
Uptake of LDL in vivo was measured from frozen liver sections with direct fluorescence after injection of DiI-labelled LDL into the marginal ear vein of WHHL rabbits 4h before sacrification. There were no differences in DiI-LDL uptake measured either as fluorescent area or as number of DiI-positive cells one month after gene transfers of LDLR or LacZ control genes in either AAV serotype (Figure 7).
Figure 7. Uptake of DiI-LDL in WHHL rabbit liver one month after AAV mediated gene transfers of rLDLR or LacZ.
5.3 SMSR DOES NOT REDUCE LESION AREA IN WHHL RABBITS (III)
5.3.1 Clinical chemistry
WHHL rabbits were transduced with 1x1011 vg of AAV2-sMSR or AAV2-EGFP after intraportal injection. Expression of the sMSR transgene was seen in all liver biopsy samples collected four weeks after gene transfer, and also in all but one animal at the end of the follow-up six months after gene transfer (Figure 8 a and b). The amount of sMSR protein in serum was analyzed with ELISA-assay, and protein was detected already one week after gene transfer with levels increasing until four weeks after gene transfer and persisting for the duration of the follow-up (Figure 8c).
No changes between the treatment groups were seen in TC and triglyceride values during the six-month follow-up (III). A transient inrcease in ALT was seen at the time of liver biopsy, otherwise liver enzymes, albumin and bilirubin levels were within normal values during the study period (III).
5.3.2 Atherosclerosis and inflammation
Atherosclerosis was evaluated after Oil Red O–staining of longitudinally opened aortas, and MOVAT staining of cross-sections at the valve level in the proximal aorta.
No significant differences were seen in either en face–lesion area (Figure 8d and e) or cross-sectional lesion area (Figure 8f and g). MOVAT-staining and immunohistochemistry for macrophages in the proximal aorta showed abundant macrophages especially in the shoulder areas of lesions ( Figure 8g and III).
Analysis of liver histology from hematoxylin-eosin stained sections revealed normal liver architecture. Some inflammatory cell infiltrates in EGFP transduced animals were seen. The number of macrophages in the livers of the different groups did not differ from each other at one month or six months after gene transfer (III). An increase was seen in the number of macrophages in the EGFP group with time, but this was not statistically significant (III).
Figure 8. Expression of sMSR mRNA in liver one month (a) and six months (b) after AAV2-sMSR gene transfer. Lanes: Mw: molecular weight marker; 1-7: liver RNA from individual animals; (-): negative control; (+): positive control. (c) sMSR protein in serum during the six-month follow-up. Quantification of en face-lesion area (d) and cross-sectional lesion area from valve level (f) six months after intraportal gene transfer of AAV2-sMSR and AAV2-control vectors. Representative images of Oil Red O-stained aortas (e). MOVAT-stain in (g) shows elastic fibers in black, ground substance in blue, muscle in red, collagen and reticular fibers in yellow and fibrin as intense red. Magnification in (g) 12.5x; scale bar 1000µm.
6 DISCUSSION
6.1 LV- rLDLR IN THE TREATMENT OF FH
Monogenic diseases were among the first candidates for gene therapy, and still remain a major interest for the development of gene therapeutics. Beneficial results have been seen for example in clinical trials of Lebers congenital amaurosis where replacement of RPE65 via a single subretinal injetion has led to a long-term expression of the transgene and improvements in visual acuity (Pennesi et al. 2018).
Even with new lipid-lowering therapies for hypercholesterolemia, such as PCSK9 inhibitors (Navarese et al. 2015), an unmet need still exists especially for hoFH patients with null-mutations who do not benefit from these new therapeutics.
Previously efficacy of the retroviral vector has been shown in WHHL rabbits for FH (Pakkanen et al. 1999). However, as retroviruses transduce only replicating cells, liver resection with thymidine kinase-ganciclovir treatment was used before gene transfer to stimulate hepatocyte proliferation to increase transduction efficiency.
This, however, is laborious and not feasible in the clinical setting. LVs are able to transduce quiescent, non-dividing cells including hepatocytes so in this study an LV expressing rLDLR under a liver-specific promoter was generated. Tissue-specific promoters are desirable since with most gene delivery routes some vector always escapes via the circulation and may transduce non-target tissues. This is especially the case with intravascular delivery methods such as intraportal injection.
The LV-rLDLR vector showed expression in HepG2 cells but not WHHL fibroblasts confirming the specificity of the LSP promoter. Increased degradation of I125 labelled LDL in HepG2 cells after LV-rLDLR transduction verified functionality
The LV-rLDLR vector showed expression in HepG2 cells but not WHHL fibroblasts confirming the specificity of the LSP promoter. Increased degradation of I125 labelled LDL in HepG2 cells after LV-rLDLR transduction verified functionality