PV-1 is recognized by the PAL-E antibody and forms complexes with NRP-1

The ability to discriminate between vessels of vascular and lymphatic origin is critical in many studies in the fields of immunology, vascular biology, pathology, and tumor biology. Several markers, including CD31, endoglin, factor VIII, and pathologische anatomie leiden-endothelium (PAL-E), have been used to identify vascular endothelium. Because of its specificity for endothelial cells of capillaries, venules, and medium-sized veins,¹  the antibody PAL-E has been the gold standard among these markers.

Despite its widespread use, the identification of the PAL-E antigen has proven difficult, and there has been marked confusion regarding the identity of the molecule recognized by PAL-E.^2-4  We identified the molecular target of PAL-E as plasmalemma vesicle-associated protein-1 (PV-1).^3,5  Recently, this was challenged by reporting neuropilin-1 (NRP-1) as the antigen for PAL-E.⁴  PV-1 is a heavily glycosylated approximately 55- to 65-kDa glycoprotein originally discovered in rat lung endothelium and subsequently found in stomatal diaphragms of endothelial caveolae and transendothelial channels.⁶  NRP-1, on the other hand, was first found as a semaphorin receptor and neuronal cell guidance molecule⁷  and later also shown to function during angiogenesis as a VEGFR-2 coreceptor binding VEGF-A₁₆₅.^8,9  This study was undertaken to elucidate the relationship between the antibody PAL-E and the proteins PV-1 and NRP-1.

For detailed information, see supplemental Methods (available on the Blood Web site; see the Supplemental Materials link at the top of the online article). The use of human material and all procedures were approved by the Ethical Board of Turku University Hospital and abided by the Declaration of Helsinki.

Frozen human tissues were stained with primary antibodies. Isotype specific secondary antibodies coupled to either Alexa Fluor-488 or Alexa Fluor-546 (Invitrogen) were used. Samples were analyzed on a LSM 510 confocal microscope (Carl Zeiss Microimaging).

For flow cytometry, HEK EBNA cells were transiently transfected (Lipofectamine 2000; Invitrogen) with pcDNA3.1 plasmids containing PV-1, NRP-1, or an empty vector. Primary antibodies were detected using FITC-conjugated secondary antibodies. The samples were analyzed on a FACSCalibur using CellQuest Pro Version 6 software (BD Biosciences) and Flowing-Software Version 2.4.7 (Perttu Terho, Center for Biotechnology; www.flowingsoftware.com).

For antibodies and staining procedures, see supplemental Methods.

For Western blot analyses, lymphocyte-depleted tonsil lysate and HEK-PV-1 transfectants were run on SDS-PAGE under nonreducing conditions, transferred to nitrocellulose membranes, and incubated with primary antibodies followed by HRP-conjugated anti–mouse IgG (Dako Denmark) or anti–goat IgG (Santa Cruz Biotechnology).

Coimmunoprecipitations were performed as described¹⁰  using fresh lymphocyte-depleted tonsil lysate or PV-1– and NRP-1–cotransfected HEK EBNA cell lysate. Magnetic beads (Dynal) were coated with 174/2 (anti–PV-1 antibody), polyclonal α-NRP-1, AK-1 (negative control), or normal goat serum (Vector Laboratories). Precipitated proteins were analyzed as described for Western blot analysis, but run under reducing conditions on SDS-PAGE.

First, we used immunofluorescence to study whether PAL-E and antibodies against PV-1 and NRP-1 would reveal differences in the expression patterns of their respective target molecules. For that purpose, we stained human peripheral lymph node, liver, thymus, brain (choroid plexus), tonsil, and heart tissues. As PV-1 is absent from normal brain endothelium and is up-regulated only after blood-brain barrier disruption,¹¹  we used choroid plexus, a part of the brain characterized by its fenestrated endothelium,¹²  expressing PV-1.

These stainings showed that PV-1 and the epitope of PAL-E colocalize perfectly in every analyzed tissue. However, double stainings of PAL-E and NRP-1 or PV-1 and NRP-1 revealed significant differences in the heart and liver (Figure 1; supplemental Videos). To confirm the nature of the stained structures, we performed double stainings of PV-1, NRP-1, and PAL-E with an antibody against the prototype vessel marker CD31/PECAM-1. Indeed, all structures stained by anti–PV-1 antibodies and PAL-E were also positive for CD31 (supplemental Figure 1). Only the antibody against NRP-1 stained some nonvascular structures in the liver.

Next, we used HEK EBNA cells transiently transfected with PV-1, NRP-1, or mock plasmids to determine the binding of the antibody PAL-E. As expected, the anti–PV-1 antibody positively stained the PV-1 transfectants, whereas 2 different anti–NRP-1 antibodies bound specifically to the NRP-1–transfected cells. The PAL-E antibody, on the other hand, positively stained PV-1 transfectants, whereas NRP-1–expressing cells remained negative (Figure 2A).

In immunoblotting (Figure 2B), PAL-E and an anti–PV-1 antibody both recognized bands of the same size (∼ 110-130 kDa) when lysates from HEK–PV-1 transfectants and tonsil were analyzed under nonreducing conditions. The anti–NRP-1 antibody recognized a specific protein band only in tonsil lysate. The absence of a signal in the HEK–PV-1 transfectants shows that the anti–NRP-1 antibody does not recognize PV-1.

To study whether PV-1 associates with NRP-1, we performed coimmunoprecipitations from fresh tonsil lysate. We found that NRP-1 can be precipitated using anti–PV-1 antibodies and immunoprecipitation with anti–NRP-1 antibodies specifically pulls down some PV-1 (Figure 2C). As the amount of PV-1 precipitated with α-NRP-1 antibodies was very low, we confirmed the physical association of PV-1 and NRP-1 using PV-1 precipitation followed by detection of NRP-1 in transiently transfected HEK-EBNA cells (Figure 2D).

Although Jaalouk et al reported that PAL-E recognizes NRP-1,⁴  the expression of NRP-1 in multiple nonvascular cells, including neuronal structures, thymic epithelial cells, T cells, and dendritic cells on one hand^13,14  and the widespread use of PAL-E as a marker for vascular endothelium more than 2 decades on the other hand, are in disagreement with this interpretation. They further mapped the PAL-E binding to the SQYSTNW motif (spanning residues 295-301) in NRP-1, which overlaps with the VEGF ligand binding site predicted earlier.¹⁵  PV-1 has no similar sequence motif. In our modelings of the 3-dimensional structure, PV-1 was predicted to be totally α-helical and PV-1 did not show any structural similarity to NRP-1. Thus, a possible structural similarity between PV-1 and NRP-1 cannot explain the false identification of NRP-1 as a PAL-E antigen.

Furthermore, the argument of Jaalouk et al,⁴  claiming that anti–PV-1 and PAL-E antibodies cannot recognize the same molecule as they fail to inhibit each other in competitive staining experiments, is invalidated by the fact that all glycoproteins carry multiple antigenic epitopes.

The function and expression of NRP-1 and VEGFR-2 in endothelial cells are closely associated.^8,16,17  In addition, a recent publication demonstrated that NRP-1 ligands, such as VEGF-A₁₆₅, contain a sequence interacting with NRP-1 and causing vascular leakage and cellular internalization.¹⁸  The authors also suggested that these functions might involve vesiculo-vacuolar organelles (VVOs) because VEGFR-2 can be detected on the luminal and abluminal surfaces as well as on the membranes of VVOs. Furthermore, VEGFR-2 was found on caveolae and the stomatal diaphragms of some VVOs,¹⁹  which are formed by PV-1.²⁰  VVOs are located at sites of neutrophil transcellular diapedesis and thought to facilitate this process.²¹  In addition, VEGF-A₁₆₅ has been shown to regulate PV-1 expression via VEGFR-2.^22,23  Finally, a recent report demonstrated a function for NRP-1 during cancer cell migration.²⁴  Based on our findings and these reports, it is probable that NRP-1 may functionally be linked to PV-1.

In conclusion, we show that the prototype endothelial marker PAL-E recognizes PV-1. However, PV-1 and NRP-1 colocalize to a very high degree and can physically interact. Hence, we connect, for the first time, these 2 proteins: one involved in leukocyte trafficking and the second in angiogenesis and cancer cell migration. This opens interesting directions for further studies in vascular biology and angiogenesis.

Contribution: J.K. designed and performed experiments, analyzed data, and wrote the manuscript; D.T. provided vital reagents; K.E. designed experiments and analyzed data; M.S. designed experiments, analyzed data, and edited the manuscript; K.A. provided vital reagents and expertise and edited the manuscript; T.S. made the structural analyses; and S.J. designed the study and experiments, analyzed data, and edited the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Sirpa Jalkanen, MediCity Research Laboratory, University of Turku, Tykistökatu 6 A, FIN-20520 Turku, Finland; e-mail: sirpa.jalkanen@utu.fi.