Transplanted endothelial cells repopulate the liver endothelium and correct the phenotype of hemophilia A mice

Transplantation of healthy cells to repair organ damage or replace deficient functions constitutes a major goal of cell therapy. However, the mechanisms by which transplanted cells engraft, proliferate, and function remain unknown. To investigate whether host liver sinusoidal endothelium could be replaced with transplanted liver sinusoidal endothelial cells, we developed an animal model of tissue replacement that utilized a genetic system to identify transplanted cells and induced host-cell perturbations to confer a proliferative advantage to transplanted cells. Under these experimental conditions, transplanted cells engrafted efficiently and proliferated to replace substantial portions of the liver endothelium. Tissue studies demonstrated that transplanted cells became integral to the liver structure and reacquired characteristic endothelial morphology. Characterization of transplanted endothelial cells by membrane markers and studies of cellular function, including synthesis and release of coagulation factor VIII, demonstrated that transplanted cells were functionally intact. Further analysis showed that repopulation of the livers of mice that model hemophilia A with healthy endothelial cells restored plasma factor VIII activity and corrected their bleeding phenotype. Our studies therefore suggest that transplantation of healthy endothelial cells should be considered for cell therapy of relevant disorders and that endothelial reconstitution with transplanted cells may offer an excellent paradigm for defining organ-specific pathophysiological mechanisms.

A major requirement of cell therapy concerns reconstitution of specific cell compartments. Recently, several studies established the possibility of replacing parenchymal liver cells with transplanted hepatic or extrahepatic donor-derived cells (1). Similarly, replacement of ECs in specific vascular beds, such as the liver, would be a profoundly significant accomplishment because ECs play central roles in liver development, organization, repair, and function. Moreover, ECs synthesize critical coagulation factors and contribute to cell-cell signaling or immunoregulation (2–5). Therefore, the ability to reconstitute the endothelium of organs with healthy cells will offer therapeutic opportunities, and organs other than the liver will also be suitable targets for this purpose, e.g., to restore cardiovascular or pulmonary functions (6, 7). At present, differences in the properties of organ-specific ECs are not fully understood. The inability of i.v.-injected pulmonary ECs to target extrapulmonary organs, e.g., the liver or spleen, and of liver sinusoidal ECs (LSECs) to target extrahepatic organs, e.g., lungs, heart, or kidneys, indicates that cell and organ type–specific differences in EC populations need to be further defined (7, 8).

To understand mechanisms in reconstitution of the hepatic endothelium, we developed robust transplantation assays with transgenic donor cells and manipulated recipients prior to cell transplantation for enhancing cell engraftment and proliferation. Similarly, model systems are needed to effectively demonstrate gain of function. Here, we addressed the role of transplanted LSECs in synthesizing and releasing the coagulation protein factor VIII (FVIII). This essential coagulation factor is lacking in hemophilia A, which has a prevalence of 1 in 10,000 males and is characterized by recurrent, spontaneous bleeding leading to disability or fatality (9). Hemophilia A is an attractive target for cell and gene therapy, as maintaining FVIII levels in the 1%–5% range decreases the risk of spontaneous bleeding. Infusion of recombinant clotting factor is effective in the short term for preventing or treating bleeding in people with hemophilia, although prevention of bleeding complications in the long term requires multiple weekly infusions, which are inconvenient, invasive, and costly. Also, generally up to 25% of hemophilia A patients develop inhibitory antibodies against FVIII clotting factors that may render further replacement therapy ineffective (10). Despite insights into genetic defects in hemophilia A, the precise organ- and cell type–specific origins of FVIII have been controversial (2, 11, 12). As LSECs produce FVIII, we addressed the therapeutic potential of liver repopulation with transplanted LSECs in hemophilia A mice. Our studies indicate that transplanted cells have potential for future clinical applications. Moreover, these studies offer suitable directions to approach restoration of additional vascular compartments with transplanted ECs.

Properties of isolated FVB/N-Tie2–GFP LSECs. To identify transplanted ECs in recipients, we used transgenic FVB/N-Tie2–GFP mice as cell donors because only ECs would express GFP under transcriptional control of the endothelial receptor tyrosine kinase promoter Tie2 (13). This GFP reporter helped in identifying LSECs during their characterization in vitro. Flow cytometric analysis, including Tie2-driven GFP expression and EC membrane markers, e.g., PECAM-1 (CD31), endoglin (CD105), Flk-1, and Tie2, indicated that isolated LSECs possessed endothelial characteristics (Figure 1, A–E, and not shown). These cells showed infrequent contaminants with 3% CD45⁺ cells (Figure 1C); within this CD45⁺ fraction, the majority were CD11b⁺CD31⁺, further indicating that cells devoid of any endothelial marker were extremely rare. Scanning electron microscopy (SEM) verified that the cell membrane of isolated LSECs contained sieve plates with grouped fenestrae, another characteristic feature of the sinusoidal liver endothelium (Figure 1, F and G) (14, 15). This established that LSECs isolated by enzymatic liver digestion followed by immunomagnetic selection with mouse endothelium–specific antibody were highly enriched (>97% ECs) and were thus suitable for our cell transplantation studies.

Figure 1

Properties of isolated FVB/N-Tie2–GFP LSECs and fate of transplanted LSECs. Flow cytometry showing GFP-positive LSECs selected with anti-LSECs after CD45-positive cells were depleted. LSECs selected showed native GFP expression in 77% (A); 96% stained for the endothelial marker CD31 (B), and only 3% stained for CD45 (C). Moreover, CD31-containing LSECs coexpressed additional endothelial markers, Flk-1 (D) and endoglin (E). (F and G) SEM showing sieve plates with fenestrae on LSEC surface (arrowheads), another feature of sinusoidal ECs. (H) Donor FVB/N-Tie2–GFP mouse liver showing GFP staining (green) in LSECs, along with a portal vein radicle (asterisk). (I–K) FVB/N mice after LSEC transplantation. (I) Two transplanted cells identified by GFP staining (green) in liver sinusoids. (J) Increased engraftment of transplanted LSECs 1 week after cell transplantation in FLP-treated mouse. (K) Significantly increased engraftment of transplanted LSECs 1 week after cell transplantation in MCT-treated mouse. (L–S) GFP fluorescence in transplanted LSECs to indicate cell proliferation 3 months after transplantation in MCT-treated mice. (L and P) Nuclear staining with DAPI (blue). (M) Kupffer cells immunostained with F480 antibody (red). (Q) Endothelial cells immunostained with CD31 antibody. (N and R) GFP immunostaining. (O and S) Merged images from all 3 panels. Original magnification, ×2,500 (F); ×12,500 (G); ×200 (H–K); ×400 (P–S); ×200 (L–O). Scale bars: 2 μm (F); 0.4 μm (G).

LSEC transplantation studies in FVB/N mice. To assess the fate of transplanted LSECs in the liver, we injected 2 × 10⁶ FVB/N-Tie2–GFP cells via the portal vein into FVB/N mice of the same genetic background. Recipient animals were subjected to interval analysis for identifying transplanted LSECs in the liver using immunostaining for GFP (Figure 1, H–K). Initially we determined whether donor LSECs would engraft efficiently in the mouse liver after transplantation. Immediately after transplantation, LSECs were distributed in liver sinusoids, and after 3–7 days, transplanted LSECs were integrated in the liver structure and engrafted in the EC layer bounding sinusoids within the liver. Analysis to demonstrate the long-term fate of cells at subsequent periods of 2 weeks and 1, 2, and 3 months showed that transplanted LSECs survived in the liver and were distinguishable from other sinusoidal liver cells throughout the duration of our studies (Figure 1, L–S). To measure what fraction of transplanted LSECs engrafted and survived in the long term, we isolated liver cells by enzymatic digestion from mice transplanted with FVB/N-Tie2–GFP LSECs followed by flow cytometry analysis of GFP-expressing cells. This demonstrated that transplanted ECs constituted less than 0.5% of the total sinusoidal liver EC fraction (see below). As the observed extent of endothelial reconstitution (<0.5%) would likely have been less than therapeutic, next we examined whether LSEC engraftment could be enhanced.

To modulate transplanted cell engraftment, we administered an engineered fibronectin-like polymer (FLP) so the principle of receptor-mediated targeting of LSECs to the liver endothelium could be addressed, as demonstrated previously for hepatic targeting of mature hepatocytes (16). Alternatively, we administered a genotoxic pyrrolizidine alkaloid, monocrotaline (MCT), which disrupts the sinusoidal endothelial barrier and facilitates incorporation of transplanted hepatocytes into the liver parenchyma (17–19). The rationale of this latter approach was based on the fact that MCT is a potent toxin to ECs and causes widespread disruption of endothelia in intact animals (20, 21). To blunt potential host immune responses against the GFP reporter in these immunocompetent FVB/N mice, our studies incorporated administration of low-dose cyclophosphamide, which in high doses can additionally deplete macrophages and disrupt endothelium (22).

We observed that in FLP-treated mice, intrahepatic retention of transplanted LSECs improved, and there was also moderate improvement in cell engraftment compared with control mice (Figure 1, I and J). On the other hand, LSEC engraftment in MCT-treated mice increased by several-fold compared with untreated controls (Figure 1K). We discovered that after only a single dose of MCT to recipients before cell transplantation, the number of healthy GFP-positive transplanted LSECs rose progressively over 3 months, indicating onset of transplanted cell proliferation (Figure 1, L–S). This was commensurate with long-lasting perturbations in the liver of LSEC recipient animals that had been exposed to MCT. Although 48 hours after MCT, prior to cell transplantation, serum alanine aminotransferase levels were mildly elevated (~3-fold), indicating acute liver injury, 1 and 2 weeks after cell transplantation, serum alanine aminotransferase levels were normal. Moreover, we did not observe any evidence for hepatic sinusoidal occlusion at the morphological level in animals studied in the short and the long term. The finding of progressive proliferation in transplanted LSECs suggested that native endothelial liver cells suffering from MCT-induced damage were at a competitive disadvantage compared with transplanted healthy cells (see below).

In previous studies lasting up to 7 days, we established that LSECs transplanted into the liver via portal vein injection survived only in the liver without redistribution to other organs, e.g., lungs, heart, spleen, kidney, or intestine (8). To establish that MCT-induced endothelial injury in additional vascular beds did not promote survival of transplanted LSECs in extrahepatic locations, we surveyed tissues for GFP transgene by PCR. We found that transplanted LSECs were still restricted to the liver in MCT-treated animals (Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/JCI32748DS1). This indicated that studies of the long-term fate of transplanted cells in the liver were appropriate.

Endothelial reconstitution by transplanted FVB/N-Tie2–GFP LSECs. To quantitate the extent of short-term and long-term endothelial reconstitution by transplanted FVB/N-Tie2–GFP LSECs, we isolated liver cells from multiple animals by enzymatic digestion followed by flow cytometric analysis. One week after LSEC transplantation, in mice without MCT pretreatment, transplanted cells constituted 0.3% ± 0.1% of the nonparenchymal liver cell fraction devoid of hepatocytes (nonparenchymal cells [NPCs]) (Figure 2A). This transplanted cell fraction was constant in mice without MCT over 1 month, indicating absence of transplanted cell proliferation (Figure 2B). However, in mice given MCT before cell transplantation, the number of transplanted LSECs was on average 4-fold greater after 1 week, indicating superior cell engraftment, and transplanted LSECs proliferated to constitute 9% of NPCs after 3 months (Figure 2, D–F). We demonstrated CD31 expression in GFP-positive transplanted cells to verify their endothelial identity (Figure 2, G–I). Cumulative analysis of these data indicated up to 30-fold increases in transplanted cell numbers over time (Table 1). Tissue analysis by GFP staining showed large numbers of GFP-positive transplanted LSECs in the liver 1 and 3 months after cell transplantation (Figure 1, L–S).

Figure 2

Flow cytometric analysis to identify reisolated transplanted FVB/N-Tie2–GFP cells as a fraction of total nonparenchymal liver cells. (A) Shown are GFP-positive cells in MCT-untreated mice, with transplanted cells constituting less than 0.5% of NPCs after 1 week and 1 month. Negative control mouse without cell transplantation is shown. (B) Upper panels show GFP-positive cells in MCT-treated mice, in which transplanted cell numbers increased progressively, reaching up to 11% of NPCs after 3 months. Lower panels show that GFP-positive cells were also CD31 positive.

Table 1

Cumulative analysis of transplanted LSEC fraction

Structural and functional integrity of transplanted LSECs in the liver. Next, we established the structural and functional integrity of reconstituted hepatic endothelium following LSEC transplantation. Transmission electron microscopy (TEM) demonstrated that transplanted LSECs were correctly integrated in the liver endothelium with the space of Disse intervening between LSECs and hepatocytes (Figure 3A). To demonstrate receptor-mediated endocytosis through scavenger receptors in transplanted LSECs, which is a physiological function of ECs, we administered to mice fluorescently labeled 1,1-dioctadecyl-3,3,3,3-tetramethylindocarbocyanine–labeled acetylated LDL (DiI-Ac-LDL) (23). Transplanted LSECs incorporated DiI-Ac-LDL 1, 3, and 7 days after cell transplantation as well as subsequently throughout the 3-month duration of our studies, including in mice with extensive proliferation of transplanted cells (Figure 3, B–F). This assay established that transplanted LSECs were functionally intact from the beginning. We again verified that transplanted LSECs possessed characteristic endothelial membrane markers, e.g., CD31 (Figure 3, G–J), and were distinct from other cells in liver sinusoids, e.g., Kupffer cells, which display the F4/80 antigen (Figure 3, K–N). Confocal microscopy demonstrated that GFP- and CD31-positive transplanted LSECs were distinct from other sinusoidal cells (Supplemental Videos 1 and 2). This meant that transplanted LSECs physiologically reconstituted the sinusoidal endothelial compartment. Absence of GFP in F4/80-stained Kupffer cells indicated that transplanted cells were not being phagocytosed, implying absence of ongoing transplanted cell destruction.

Figure 3

Integrity of transplanted LSECs in the liver. (A) TEM after GFP immunostaining showing integration of transplanted LSECs after 1 week (arrows) with characteristic layering of hepatic sinusoidal endothelium (arrowheads). Asterisk indicates lipid-laden stellate cells in adjacent area. (B–F) Analysis of DiI-Ac-LDL uptake (red) in GFP-positive LSECs (green) producing yellow color in cells containing both. Shown are transplanted GFP-positive LSECs in FVB/N mice containing green, yellow, or red domains due to DiI-Ac-LDL uptake in cells 3 months (E) after cell transplantation. (G–J) Combined staining for GFP (H) and CD31 (I) verifying transplanted cells expressed both markers (yellow). (K–N) Costaining for F4/80 (L, red) and GFP (M, green) showing absence of F4/80 immunostaining in transplanted LSECs, indicating distinction from Kupffer cells. (G and K) Nuclei stained with DAPI (blue). Original magnification, ×2,700 (A); ×200 (B); ×400 (C–F); ×630 (G–N). Scale bars: 5 μm (A); 10 μm (G–N).

MCT-induced perturbations in LSECs. Previously, we established that MCT promptly disrupted liver sinusoidal endothelium, which was demonstrable at the morphological level, as well as through impaired clearance of endothelial ligands, e.g., serum hyaluronic acid, within 24–48 hours (17, 18). If MCT perturbed the liver in the short term as well as the long term, we should have been able to demonstrate relevant changes in LSECs. For instance, changes in cell signaling molecules regulating endothelial remodeling or angiogenesis should have occurred shortly after MCT treatment. Our studies in animals treated several hours previously with MCT verified such a possibility. We observed prompt increases in the hepatic expression of MMP-9, VEGF receptor-2, and the chemokine receptor CXCR4, specific for stromal cell–derived factor-1 (SDF-1), whereas expression of VEGF or SDF-1 in the liver was unchanged (Figure 4). As MMPs, CXCR4, VEGF, and SDF-1 mediate processes involved in endothelial remodeling during transplanted cell engraftment in the liver (24, 25), these findings were in agreement with activation of engraftment-promoting events in MCT-treated animals. In this situation, recruitment of cell-specific receptors would help targeting of transplanted cells to the endothelium whereas matrix-degrading and matrix-reconstituting MMP activities will benefit permanent integration of transplanted cells in the liver structure.

Figure 4

RT-PCR analysis of total liver RNA from FVB/N mice treated with MCT. RT-PCR for several genes (MMP-9, CXCR4, SDF-1, VEGFR2, and VEGF) was performed in mice sacrificed at various times after MCT. For gene expression, liver RNA from 2 mice per condition was analyzed. The data showed increased expression of MMP-9, CXCR4, and VEGFR2.

Similarly, MCT should have impaired survival and/or proliferation capacity in native ECs following genotoxicity, which is a known feature of pyrrolizidine alkaloids (20). To analyze changes in survival of native LSECs, we examined FVB/N mice 2–4 weeks after MCT. We chose these times to ensure that acute injury following MCT will have resolved and that ECs will not have been severely perturbed by MCT. The viability of freshly isolated LSECs 2 weeks after MCT was impaired: 74% ± 4% versus 92% ± 3% in LSECs from untreated control mice, P < 0.001. In freshly isolated LSECs from healthy donors, 5% were undergoing apoptosis as shown by TUNEL, whereas 23% freshly isolated LSECs from donors 1 month after MCT were positive for TUNEL, P < 0.001 (Supplemental Figure 2).

Also, we performed studies in MCT-pretreated FVB/N recipients with transplantation of LSECs from FVB/N-Tie2–GFP donors treated 2 weeks previously with MCT. These MCT-treated donor cells engrafted in the liver and displayed suitable endothelial markers (Figure 5, A and B). However, the capacity of these cells to proliferate was markedly impaired (Figure 5C). Quantitative analysis of reisolated transplanted cells showed that 1 week after cell transplantation, engraftment of FVB/N-Tie2–GFP LSECs from MCT-treated donors was at least 10-fold lower than healthy LSEC donors, 0.1% vs. 1.3% transplanted LSECs in NPCs by flow cytometry, P < 0.001 (Figure 5D and Table 2). Similarly, 1 month after cell transplantation, 10-fold fewer transplanted LSECs were recovered from MCT-treated donors compared with healthy donors. In MCT-treated animals in which no cells were transplanted, widespread apoptosis was observed in sinusoidal liver cells, which was different from animals in which transplanted LSECs were undergoing extensive proliferation (Supplemental Figure 3, D–F). To identify cell fractions undergoing cycling, we performed Ki67 staining. In normal mouse liver, ECs did not express Ki67. In contrast, in MCT-treated mice in which transplanted cells were undergoing proliferation, liver sinusoidal cells expressed Ki67 extensively 1 week as well as 1 month after cell transplantation (Supplemental Figure 4, F–H). Therefore, we concluded that MCT impaired the capacity of native LSECs to survive and to proliferate, thereby offering advantages to transplanted LSECs in the liver.

Figure 5

Perturbations induced by MCT in native LSECs. Analysis of liver from MCT-pretreated FVB/N recipient mice after transplantation of LSECs from FVB/N-Tie2–GFP donor mice that had also been treated with MCT 2 weeks before cells were isolated for transplantation. (A) Tissue sections were immunostained with GFP (green) and F4/80 (red) antibodies 1 week after cell transplantation. (B) Higher magnification view of GFP and CD31 staining in a section from mouse 1 week after cell transplantation. Arrows indicate GFP-positive transplanted LSECs displaying CD31. (C) Immunostaining to identify transplanted LSECs with GFP (green) along with CD31 endothelial marker (red) 1 month after cell transplantation. (D) FACS analysis of MCT-treated FVB/N-Tie2–GFP transplanted LSECs recovered 1 week and 1 month after transplantation. The identities of CD31⁺/GFP LSECs for typical endothelial markers, endoglin and Flk-1, and the myeloid marker CD11b are shown in the last 6 plots of LSECs isolated from mice 1 week after cell transplantation. Nuclei stained with DAPI (blue). Original magnification, ×200 (A, C); ×630 (B). Scale bar: 10 μm (B).

Table 2

Recovery of transplanted cells

These mechanisms provided suitable experimental systems for addressing repopulation of the endothelial compartment although potential additional advantages offered by the incorporation of cyclophosphamide required consideration. In FVB/N mice treated with cyclophosphamide alone, transplanted LSECs did not proliferate, indicating that by itself, cyclophosphamide was not responsible for this effect (Figure 2B). Short-term in vitro assays showed that cyclophosphamide did not enhance MCT-induced cytotoxicity in LSECs (Supplemental Figure 5). However, as cyclophosphamide may not be toxic to cultured LSECs without hepatocytes (26), we could not exclude the possibility of different in vivo effects.

Function of transplanted healthy LSECs in NOD/SCID hemophilia A mice. As plasma FVIII activity is undetectable in hemophilia A mice, we hypothesized that transplanted LSECs producing and releasing sufficient amounts of FVIII in blood would result in therapeutic correction of hemophilia A. These FVIII knockout mice constitute an excellent model for addressing cell therapy in hemophilia A (27–29). We transplanted healthy FVIII-expressing LSECs from FVB/N-Tie2–GFP donors via the portal vein into hemophilia A mice tat were backcrossed 10 times with T and B cell–deficient NOD.CB17-Prkdc^scid/J mice. This was done to eliminate the possibility of immune responses against FVIII secreted by transplanted cells as well as against transplanted cells themselves. We administered MCT to hemophilia A mice (n = 31) from 1 to 2 days before cell transplantation. To perform surgery in these mice, we administered 4 units of recombinant human FVIII once, which exhibits terminal half-life of 4.1 hours in mice (30).

Two mice died before reaching the 1-month time point and were lost to analysis. Subsequently, in 1 set of studies, we examined tissues and plasma from 16 animals 1 month after LSEC transplantation, although plasma from 1 of these animals was lost during sampling. RT-PCR demonstrated FVIII mRNA in the liver of virtually all mice (Figure 6A). Again, we identified transplanted LSECs with GFP and CD31, distinct from F4/80-reactive Kupffer cells (Figure 6, B and C, and Figure 3, G–N). The relative abundance of hepatic FVIII mRNAs correlated with the prevalence of transplanted LSECs in the liver of hemophilic mice. We demonstrated plasma FVIII activity with a sensitive fluorogenic thrombin generation test (31), arbitrarily using a stringent cutoff of more than 10% plasma FVIII activity to define therapeutic correction. Significant correlation was found between the overall extent of liver repopulation with transplanted LSECs and plasma FVIII activity. Our criterion of therapeutic correction indicated that transplantation of healthy LSECs corrected hemophilia A in 13 of 15 mice (87%). Mean FVIII activity in these mice was 19% ± 4% of normal plasma activity (range: 14%–25%) (Figure 6D). In 2 mice judged to be without therapeutic correction, FVIII activity was either undetectable or only 7%. Transplanted cells were less prevalent in the liver of these 2 mice compared with mice showing therapeutic correction.

Figure 6

Therapeutic efficacy of transplanted LSECs in NOD/SCID hemophilia A mice. (A) RT-PCR analysis of total liver RNA showing FVIII mRNA in virtually all treated mice (lanes 2–13). The samples were analyzed for β-actin in parallel to verify RNA integrity. Lane 1, negative control with PCR mix alone; lane 14, healthy mouse liver RNA as positive control. (B) Immunostaining of liver showing transplanted FVB/N-Tie2–GFP LSECs in NOD/SCID hemophilia A mouse (green) with coexpression of CD31 endothelial marker (red in native cells; yellow in transplanted cells). (C) GFP-positive transplanted cells (green) distinct from F4/80-expressing Kupffer cells (red). (D) Plasma FVIII activity in NOD/SCID hemophilia A mice 1 month after transplantation of FVB/N-Tie2–GFP LSECs with therapeutic correction requiring more than 10% plasma FVIII. Original magnification, ×200 (B and C).

In hemophilia A mice subjected to LSEC transplantation, FVIII antigen appeared in the blood (Figure 7A), corresponding to increased plasma FVIII activity. We did not detect antibodies against FVIII in hemophilia A mice treated with healthy LSECs. As NOD/SCID mice lack the intrinsic ability to mount antibody responses due to lack of B (and T) cells, this was expected. To verify that plasma FVIII levels increased progressively in hemophilia A mice during proliferation of transplanted LSECs, we obtained blood samples serially in another set of mice over up to 2 months (Figure 7B). Plasma FVIII levels increased over time in relationship with increased liver repopulation with transplanted cells. This analysis showed that plasma FVIII activity increased from 19% ± 11% after 2 weeks to 36% ± 16% after 2 months following cell transplantation, although the difference did not achieve statistical significance due to animal-to-animal variabilities. Nonetheless, plasma FVIII activity was not lost in any of the hemophilia A mice, which indicated persistent function of transplanted cells.

Figure 7

FVIII antigen in LSEC-transplanted hemophilia A mice and serial analysis of FVIII activity in cell recipients. (A) Hemophilia A mice 1 month after transplantation of FVB/N-Tie2–GFP LSECs with therapeutic correction. Neg Ctr, untreated hemophilia A mice. Each bar represents FVIII antigen level from an individual mouse after cell transplantation. (B) Plasma FVIII activity in hemophilia A mice after cell transplantation showing serial bleeding from 7 mice starting 2 weeks after surgery over a period of 60 days with therapeutic correction requiring more than 10% plasma FVIII. Mouse indicated as sham received vehicle alone.

Restoration of plasma FVIII activity should have corrected the bleeding phenotype in hemophilia A mice. This was examined in another set of studies. In hemophilia A mice 2 months following LSEC transplantation, tail cut–induced bleeding stopped after 15–120 minutes, and 6 of 6 hemophilia A mice survived (100%). In contrast, in untreated control hemophilia A mice, bleeding continued and proved fatal.

We expected that in hemophilia A mice, other coagulation factors and cofactors would be expressed normally, e.g., vWF or coagulation factor IX (FIX), since no other gene was disrupted. This was demonstrated by the presence of vWF and FIX mRNAs (Figure 8A) as well as vWF protein in LSECs from hemophilia A mice (Figure 8, B and C, and Supplemental Video 3 showing confocal microscopy). Similarly, hemophilia A mice subjected to LSEC transplantation also expressed vWF (Figure 8D). It was difficult to discern whether restoration of FVIII expression in these animals had counterregulatory effects on vWF expression. Finally, to determine whether ECs in other organs could offer opportunities to reconstitute FVIII deficiency, we studied FVIII expression in several organs (Figure 8E). FVIII mRNA was present in spleen, lungs, and bone marrow, although at lower levels than in liver. To entertain the possibility of cell transplantation studies, we isolated ECs from the mouse spleen by immunomagnetic cell sorting. However, the yield of splenic ECs was extremely limited, and we could isolate less than 5 × 10⁵ ECs from the spleen, which restricted in vivo studies.

Figure 8

Expression of coagulation factors in ECs. (A) RT-PCR of total liver RNA for FVIII, FIX, and vWF. Lane 1, liver from healthy FVB/N mouse; lane 2, liver from hemophilia A mouse; lane 3, LSECs from healthy FVB/N mouse; lane 4, hepatocytes from healthy FVB/N mouse; and lane 5, PCR mix alone. (B and C) Immunostaining of NOD/SCID hemophilia A mouse liver showing coexpression of CD31 endothelial marker (red) and vWF (green) in liver sinusoids; cells expressing both appear yellow. C shows enlargement of the boxed area in B. Nuclei were stained with DAPI (blue). In B, confocal microscopy demonstrates vWF expression in LSECs costained for CD31. Scale bar: 24 μm (B). Original magnification, ×630 (C). (D) RT-PCR of total liver RNA showing vWF mRNA expression. Lane 1, FVB/N donor LSECs; lane 2, hemophilia A mouse liver; lane 3, FVB/N mouse liver; lanes 4–7, liver of hemophilic mice after LSEC transplantation. (E) RT-PCR of total RNA for FVIII expression. Lanes 1–4, FVB/N mouse liver, bone marrow, spleen, and lung, respectively; lane 5, hemophilia A mouse liver; lane 6, FVB/N LSECs; lane 7, FVB/N ECs from spleen; and lane 8, mix alone. Expression of β-actin verified RNA integrity in A, D, and E.

Although mature ECs reconstituted hepatic sinusoidal endothelium in our studies, additional cell types, including circulating mononuclear cell–derived or other endothelial progenitor cells, may home to areas of endothelial damage and contribute to vasculogenesis or neoangiogenesis (32–35) and may improve outcomes in animals with liver disease (36). In our studies, donor cells were mature LSECs, and contamination with other cell types was essentially of no significance. Under suitable growth factor stimulation conditions, primary mature LSECs can proliferate in vitro, and as now shown in our studies, also in vivo. Judging from our experience in isolating ECs from the spleen, organs other than the liver may not necessarily provide more ECs. Although isolation of primary cells from human livers requires significant expertise, donor livers are processed routinely to isolate hepatocytes and other liver cells. We considered that transplantation of 2 × 10⁶ LSECs in mice represented 10% of the endothelial compartment and 5% of the NPC fraction, assuming that LSECs, NPCs, and hepatocytes constituted 20%, 40%, and 60%, respectively, of 1 × 10⁸ total liver cells. In principle, a single donor human liver should generate sufficient numbers of LSECs to treat multiple patients if only 10% of the LSECs were needed for transplantation.

Reisolation of up to 0.5% transplanted LSECs in NPCs from mice without MCT preconditioning suggested that only 10% of the transplanted LSECs engrafted, which was generally similar to the engraftment efficiency of primary hepatocytes in the liver (37). On the other hand, pretreatment of mice with FLP and MCT, especially MCT, significantly enhanced LSEC engraftment, also resembling the effect of MCT on the engraftment of transplanted hepatocytes, which too is facilitated by the disruption of the sinusoidal endothelial barrier (17–19). The mechanism by which FLP works is quite different because MCT damages cells, whereas FLP engages extracellular matrix receptors without causing cell damage (16). Subsequently, transplanted LSECs proliferated in MCT-pretreated mice due to impairment in the survival and proliferation capacity of native LSECs, as shown by our studies, which were in agreement with the propensity of MCT to induce persistent cell-cycle arrest, such as through G2/M block, in ECs (38, 39). Consequently, the extent of liver repopulation with LSECs reached 10%–20% of the endothelial mass, judging from the analysis of cell fractions in our studies. In contrast, MCT did not affect hepatocytes in this manner, unless hepatic injury was redirected by additional drug toxicities or accumulation of toxic MCT intermediates through activation of cytochrome P450–induced metabolism of MCT in hepatocytes (18). The principle of selective injury in native cells can possibly be applied to repopulation of EC compartments in other vascular beds and assessment the potential of cell-cell interactions in various conditions.

To consider clinical applications, substitutes for MCT will be necessary, since this substance has oncogenic potential. As cyclophosphamide and doxorubicin, which are already in clinical use, were also effective for disrupting the hepatic endothelial barrier to improve engraftment of transplanted hepatocytes (22, 40), potential alternative approaches should be feasible. Identification of suitable changes in the liver that promote proliferation of transplanted LSECs should provide impetus for further research into suitable drugs capable of reproducing these mechanisms.

It was particularly noteworthy that transplanted LSECs maintained physiological function in the liver. Our studies unequivocally established that LSECs are the major source of FVIII. For targeted cell and gene therapy in hemophilia A, identification of the cell type(s) capable of producing FVIII will be of enormous value. Although whole liver transplantation in people with hemophilia A promptly resulted in the restoration of plasma FVIII activity, supporting a role of the liver in FVIII synthesis (41), the specific cell type(s) responsible for producing FVIII had remained elusive because FVIII mRNA was expressed in hepatocytes, LSECs, and macrophages (2, 12, 29). The cell type–specific origin of FVIII was not resolved by further studies since transplantation of healthy LSECs (29), fetal spleen containing vascular endothelium (28), and hepatocytes (42) rescued hemophilia A mice. FVIII mRNA is particularly well expressed in hepatocytes and LSECs, although quantitative assays demonstrated greater mRNA expression in the latter (2). We consider that LSECs are the major source of FVIII protein, since LSEC transplantation in hemophilia A mice restored plasma FVIII activity levels to 14%–25% of normal, which correlated to the extent by which the hepatic endothelium was replaced by transplanted LSECs. It should be noteworthy that these levels of plasma FVIII activity eliminate spontaneous bleeding in hemophilia A.

No other major cell type capable of producing FVIII had been transplanted in our studies. While plasma FVIII may originate from additional sources, e.g., peripheral mononuclear cell derived–blood outgrowth ECs (11) or lung microvascular ECs (43), LSECs possess the intrinsic advantage of simultaneously synthesizing, sequestering, and releasing vWF, which stabilizes and protects FVIII against premature degradation or cellular incorporation (44). Moreover, LSECs play roles in immunoregulation, which could potentially be harnessed to avoid deleterious immune responses against FVIII (3). Therefore, the ability of transplanted healthy LSECs to treat FVIII deficiency in hemophilia A should offer exciting new therapeutic approaches.

Development of clinical applications with transplanted ECs for vascular reconstitution should be facilitated by the identification of effective sources of donor cells. Since the supply of donor human organs is limited, whether donor cells could be derived from alternative sources, including stem cells, will need consideration. The possibility of using genetically modified autologous stem cells after appropriate expansion in vitro could help avoid allograft-related issues.

View Supplemental data

View Supplemental video 1

View Supplemental video 2

View Supplemental video 3

This work was supported in part by NIH grants R01 DK071111, M01 RR12248, P30 DK41296, P30 CA13330, and CA06576. We are grateful to Chaoying Zhang, Gertrude Arason, and Simone Merlin for technical assistance. A. Follenzi is a recipient of a Liver Scholar Award from the American Liver Foundation/American Association for Studies of Liver Diseases. We thank R. Sarkar and L. Naldini for critical reviews of the manuscript and for helpful discussions.

Address correspondence to: Sanjeev Gupta, Albert Einstein College of Medicine, Ullmann Building, Room 625, 1300 Morris Park Avenue, Bronx, New York 10461, USA. Phone: (718) 430-2098; Fax: (718) 430-8975; E-mail: sanjvgupta@pol.net.

Nonstandard abbreviations used: DiI-Ac-LDL, 1,1-dioctadecyl-3,3,3,3-tetramethylindocarbocyanine–labeled acetylated LDL; FIX, coagulation factor IX; FLP, fibronectin-like polymer; FVIII, coagulation factor VIII; LSEC, liver sinusoidal EC; MCT, monocrotaline; NPC, nonparenchymal cell; SDF-1, stromal cell–derived factor-1; SEM, scanning electron microscopy; TEM, transmission electron microscopy.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J. Clin. Invest.118:935–945 (2008). doi:10.1172/JCI32748

1. Gupta, S., Inada, M., Joseph, B., Kumaran, V., Benten, D. 2004. Emerging insights into liver-directed cell therapy for genetic and acquired disorders. Transpl. Immunol. 12:289-302.
   View this article via: CrossRef PubMed Google Scholar
2. Do, H., Healey, J.F., Waller, E.K., Lollar, P. 1999. Expression of factor VIII by murine liver sinusoidal endothelial cells. J. Biol. Chem. 274:19587-19592.
   View this article via: CrossRef PubMed Google Scholar
3. Knolle, P.A., Limmer, A. 2003. Control of immune responses by savenger liver endothelial cells. Swiss Med. Wkly. 133:501-506.
   View this article via: PubMed Google Scholar
4. Rafii, S., Lyden, D. 2003. Therapeutic stem and progenitor cell transplantation for organ vascularization and regeneration. Nat. Med. 9:702-712.
   View this article via: CrossRef PubMed Google Scholar
5. Lamalice, L., Le Boeuf, F., Huot, J. 2007. Endothelial cell migration during angiogenesis. Circ. Res. 100:782-794.
   View this article via: CrossRef PubMed Google Scholar
6. Griese, D.P., et al. 2003. Isolation and transplantation of autologous circulating endothelial cells into denuded vessels and prosthetic grafts: implications for cell-based vascular therapy. Circulation. 108:2710-2715.
   View this article via: CrossRef PubMed Google Scholar
7. Ewing, P., et al. 2003. Isolation and transplantation of allogeneic pulmonary endothelium derived from GFP transgenic mice. J. Immunol. Methods. 283:307-315.
   View this article via: CrossRef PubMed Google Scholar
8. Benten, D., et al. 2005. Hepatic targeting of transplanted liver sinusoidal endothelial cells in intact mice. Hepatology. 42:140-148.
   View this article via: CrossRef PubMed Google Scholar
9. Bolton-Maggs, P.H., Pasi, K.J. 2003. Haemophilias A and B. Lancet. 361:1801-1809.
   View this article via: CrossRef PubMed Google Scholar
10. Teitel, J., et al. 2007. A systematic approach to controlling problem bleeds in patients with severe congenital haemophilia A and high-titre inhibitors. Haemophilia. 13:256-263.
    View this article via: CrossRef PubMed Google Scholar
11. Lin, Y., et al. 2002. Use of blood outgrowth endothelial cells for gene therapy for hemophilia A. Blood. 99:457-462.
    View this article via: CrossRef PubMed Google Scholar
12. Hollestelle, M.J., et al. 2001. Tissue distribution of factor VIII gene expression in vivo--a closer look. Thromb. Haemost. 86:855-861.
    View this article via: PubMed Google Scholar
13. Motoike, T., et al. 2000. Universal GFP reporter for the study of vascular development. Genesis. 28:75-81.
    View this article via: CrossRef PubMed Google Scholar
14. DeLeve, L.D., Wang, X., McCuskey, M.K., McCuskey, R.S. 2006. Rat liver endothelial cells isolated by anti-CD31 immunomagnetic separation lack fenestrae and sieve plates. Am. J. Physiol. Gastrointest. Liver Physiol. 291:G1187-G1189.
    View this article via: CrossRef PubMed Google Scholar
15. Nagai, T., et al. 2004. Actin filaments around endothelial fenestrae in rat hepatic sinusoidal endothelial cells. Med. Electron Microsc. 37:252-255.
    View this article via: CrossRef PubMed Google Scholar
16. Kumaran, V., Joseph, B., Benten, D., Gupta, S. 2005. Integrin and extracellular matrix interactions regulate engraftment of transplanted hepatocytes in the rat liver. Gastroenterology. 129:1643-1653.
    View this article via: CrossRef PubMed Google Scholar
17. Joseph, B., et al. 2006. Monocrotaline promotes transplanted cell engraftment and advances liver repopulation in rats via liver conditioning. Hepatology. 44:1411-1420.
    View this article via: CrossRef PubMed Google Scholar
18. Wu, Y.M., Joseph, B., Berishvili, E., Kumaran, V., Gupta, S. 2008. Hepatocyte transplantation and drug-induced perturbations in liver cell compartments. Hepatology. 47:279-287.
    View this article via: CrossRef PubMed Google Scholar
19. Witek, R.P., Fisher, S.H., Petersen, B.E. 2005. Monocrotaline, an alternative to retrorsine-based hepatocyte transplantation in rodents. Cell Transplant. 14:41-47.
    View this article via: PubMed Google Scholar
20. Copple, B.L., Banes, A., Ganey, P.E., Roth, R.A. 2002. Endothelial cell injury and fibrin deposition in rat liver after monocrotaline exposure. Toxicol. Sci. 65:309-318.
    View this article via: CrossRef PubMed Google Scholar
21. Deleve, L.D., et al. 2003. Sinusoidal obstruction syndrome (veno-occlusive disease) in the rat is prevented by matrix metalloproteinase inhibition. Gastroenterology. 125:882-890.
    View this article via: CrossRef PubMed Google Scholar
22. Malhi, H., et al. 2002. Cyclophosphamide disrupts hepatic sinusoidal endothelium and improves transplanted cell engraftment in rat liver. Hepatology. 36:112-121.
    View this article via: CrossRef PubMed Google Scholar
23. Bailey, A.S., et al. 2004. Transplanted adult hematopoietic stems cells differentiate into functional endothelial cells. Blood. 103:13-19.
    View this article via: CrossRef PubMed Google Scholar
24. Grunewald, M., et al. 2006. VEGF-induced adult neovascularization: recruitment, retention, and role of accessory cells. Cell. 124:175-189.
    View this article via: CrossRef PubMed Google Scholar
25. Kollet, O., et al. 2003. HGF, SDF-1, and MMP-9 are involved in stress-induced human CD34+ stem cell recruitment to the liver. J. Clin. Invest. 112:160-169.
    View this article via: JCI CrossRef PubMed Google Scholar
26. DeLeve, L.D. 1996. Cellular target of cyclophosphamide toxicity in the murine liver: role of glutathione and site of metabolic activation. Hepatology. 24:830-837.
    View this article via: CrossRef PubMed Google Scholar
27. Bi, L., et al. 1995. Targeted disruption of the mouse factor VIII gene produces a model of haemophilia A. Nat. Genet. 10:119-121.
    View this article via: CrossRef PubMed Google Scholar
28. Aronovich, A., et al. 2006. Correction of hemophilia as a proof of concept for treatment of monogenic diseases by fetal spleen transplantation. Proc. Natl. Acad. Sci. U. S. A. 103:19075-19080.
    View this article via: CrossRef PubMed Google Scholar
29. Kumaran, V., et al. 2005. Transplantation of endothelial cells corrects the phenotype in hemophilia A mice. J. Thromb. Haemost. 3:2022-2031.
    View this article via: CrossRef PubMed Google Scholar
30. Mordenti, J., et al. 1996. Pharmacokinetics and interspecies scaling of recombinant human factor VIII. Toxicol. Appl. Pharmacol. 136:75-78.
    View this article via: CrossRef PubMed Google Scholar
31. James, P.D., et al. 2005. Aminoglycoside suppression of nonsense mutations in severe hemophilia. Blood. 106:3043-3048.
    View this article via: CrossRef PubMed Google Scholar
32. De Palma, M., et al. 2005. Tie2 identifies a hematopoietic lineage of proangiogenic monocytes required for tumor vessel formation and a mesenchymal population of pericyte progenitors. Cancer Cell. 8:211-226.
    View this article via: CrossRef PubMed Google Scholar
33. Shaked, Y., et al. 2005. Genetic heterogeneity of the vasculogenic phenotype parallels angiogenesis; Implications for cellular surrogate marker analysis of antiangiogenesis. Cancer Cell. 7:101-111.
    View this article via: PubMed Google Scholar
34. Maruyama, K., et al. 2005. Inflammation-induced lymphangiogenesis in the cornea arises from CD11b-positive macrophages. J. Clin. Invest. 115:2363-2372.
    View this article via: JCI CrossRef PubMed Google Scholar
35. Lewis, C.E., De Palma, M., Naldini, L. 2007. Tie2-expressing monocytes and tumor angiogenesis: regulation by hypoxia and angiopoietin-2. Cancer Res. 67:8429-8432.
    View this article via: CrossRef PubMed Google Scholar
36. Taniguchi, E., et al. 2006. Endothelial progenitor cell transplantation improves the survival following liver injury in mice. Gastroenterology. 130:521-531.
    View this article via: CrossRef PubMed Google Scholar
37. Gupta, S., Malhi, H., Gagandeep, S., Novikoff, P. 1999. Liver repopulation with hepatocyte transplantation: new avenues for gene and cell therapy. J. Gene Med. 1:386-392.
    View this article via: CrossRef PubMed Google Scholar
38. Wilson, D.W., Lame, M.W., Dunston, S.K., Segall, H.J. 2000. DNA damage cell checkpoint activities are altered in monocrotaline pyrrole-induced cell cycle arrest in human pulmonary artery endothelial cells. Toxicol. Appl. Pharmacol. 166:69-80.
    View this article via: CrossRef PubMed Google Scholar
39. Shah, M., Patel, K., Sehgal, P.B. 2005. Monocrotaline pyrrole-induced endothelial cell megalocytosis involves a Golgi blockade mechanism. Am. J. Physiol. Cell Physiol. 288:C850-C862.
    View this article via: CrossRef PubMed Google Scholar
40. Kim, K.S., Joseph, B., Inada, M., Gupta, S. 2005. Regulation of hepatocyte engraftment and proliferation after cytotoxic drug-induced perturbation of the rat liver. Transplantation. 80:653-659.
    View this article via: CrossRef PubMed Google Scholar
41. Bontempo, F.A., et al. 1987. Liver transplantation in hemophilia A. Blood. 69:1721-1724.
    View this article via: PubMed Google Scholar
42. Ohashi, K., et al. 2005. Liver tissue engineering at extrahepatic sites in mice as a potential new therapy for genetic liver diseases. Hepatology. 41:132-140.
    View this article via: CrossRef PubMed Google Scholar
43. Jacquemin, M., et al. 2006. FVIII production by human lung microvascular endothelial cells. Blood. 108:515-517.
    View this article via: CrossRef PubMed Google Scholar
44. Dasgupta, S., et al. 2007. VWF protects FVIII from endocytosis by dendritic cells and subsequent presentation to immune effectors. Blood. 109:610-612.
    View this article via: CrossRef PubMed Google Scholar
45. Chao, H., Monahan, P.E., Liu, Y., Samulski, R.J., Walsh, C.E. 2001. Sustained and complete phenotype correction of hemophilia B mice following intramuscular injection of AAV1 serotype vectors. Mol. Ther. 4:217-222.
    View this article via: CrossRef PubMed Google Scholar
46. Fields, P.A., et al. 2001. Risk and prevention of anti-factor IX formation in AAV-mediated gene transfer in the context of a large deletion of F9. Mol. Ther. 4:201-210.
    View this article via: CrossRef PubMed Google Scholar
47. Novikoff, P.M., et al. 1996. Three-dimensional organization of rat hepatocyte cytoskeleton: relation to the asialoglycoprotein endocytosis pathway. J. Cell Sci. 109:21-32.
    View this article via: PubMed Google Scholar
48. Chao, H., et al. 2003. Phenotype correction of hemophilia A mice by spliceosome-mediated RNA trans-splicing. Nat. Med. 9:1015-1019.
    View this article via: CrossRef PubMed Google Scholar
49. Wu, T., Handa, J.T., Gottsch, J.D. 2005. Light-induced oxidative stress in choroidal endothelial cells in mice. Invest. Ophthalmol. Vis. Sci. 46:1117-1123.
    View this article via: CrossRef PubMed Google Scholar
50. Chen, S.D., et al. 2005. Pravastatin attenuates ceramide-induced cytotoxicity in mouse cerebral endothelial cells with HIF-1 activation and VEGF upregulation. Ann. N. Y. Acad. Sci. 1042:357-364.
    View this article via: CrossRef PubMed Google Scholar
51. Brekken, R.A., et al. 2003. Enhanced growth of tumors in SPARC null mice is associated with changes in the ECM. J. Clin. Invest. 111:487-495.
    View this article via: JCI CrossRef PubMed Google Scholar
52. Wang, F.S., et al. 2003. Population survey of CCR5 delta32, CCR5 m303, CCR2b 64I, and SDF1 3’A allele frequencies in indigenous Chinese healthy individuals, and in HIV-1-infected and HIV-1-uninfected individuals in HIV-1 risk groups. J. Acquir. Immune Defic. Syndr. 32:124-130.
    View this article via: PubMed Google Scholar
53. Molyneaux, K.A., et al. 2003. The chemokine SDF1/CXCL12 and its receptor CXCR4 regulate mouse germ cell migration and survival. Development. 130:4279-4286.
    View this article via: CrossRef PubMed Google Scholar
54. McIntosh, J.H., Owens, D., Lee, C.A., Raut, S., Barrowcliffe, T.W. 2003. A modified thrombin generation test for the measurement of factor VIII concentrates. J. Thromb. Haemost. 1:1005-1011.
    View this article via: CrossRef PubMed Google Scholar
55. Hemker, H.C. 1999. Endogenous thrombin potential. In Laboratory techiques in thrombosis: a manual. J. Jespersen, R.M. Bertina, and F. Haverkate, editors. Kluwer Academic Publishers. Dordrecht, The Netherlands. 63–79.
56. Kasper, C.K., et al. 1975. Proceedings: A more uniform measurement of factor VIII inhibitors. Thromb. Diath. Haemorrh. 34:612.
    View this article via: PubMed Google Scholar
57. Verbruggen, B., et al. 1995. The Nijmegen modification of the Bethesda assay for factor VIII:C inhibitors: improved specificity and reliability. Thromb. Haemost. 73:247-251.