Platelets and erythrocyte-bound platelets bind infectious HIV-1 in plasma of chronically infected patients

Abstract

Chronic HIV-1 infection is associated with persistent viremia in most patients, but it remains unclear how free virus may survive the potential hostile effects of plasma. We investigated whether sites might exist on the surfaces of circulating blood cells for protection of infectious HIV-1 particles. Red blood cells (RBC) either from blood of uninfected normal individuals, or from blood obtained without EDTA from chronically infected HIV-1 patients, invariably contained a small number of RBC having attached platelets as determined by flow cytometry, light microscopy, and immunofluorescence microscopy. After mixing normal RBC with platelet-rich plasma, discrete populations of RBC, platelets, and complexes of platelets attached to RBC were purified by fluorescence-activated cell sorting. Upon incubation of purified cells or platelets with HIV-1 followed by washing and co-incubation with CD4-positive peripheral blood mononuclear cells (PBMC), platelets, and platelet-RBC complexes, but not platelet-free RBC, caused infection of PBMC. Infection was prevented by pre-treating the platelet-RBC complexes with EDTA. Plasma and RBC (comprising a RBC/platelet-RBC mixture) from chronically infected patients with low viral loads were also co-incubated with PBMC ex vivo to determine the presence of infectious HIV-1. All freshly isolated plasmas from the HIV-1-infected donors, obtained in the absence of anticoagulant, were noninfectious. Interestingly, the RBC from most of the patients caused cell-cell infection of PBMC that was prevented by stripping the RBC with EDTA. A monoclonal antibody to DC-SIGN partially inhibited cell-cell HIV-1 infection of PBMC by normal RBC pre-incubated with platelets and HIV-1. We conclude: (a) platelet-free EDTA-free plasma from chronically infected HIV-1 patients, although containing viral RNA, is an environment that lacks detectable infectious HIV-1; (b) platelets and platelet-RBC complexes, but not purified RBC, bind infectious HIV-1; (c) DC-SIGN, and possibly other C-type lectins, may represent binding sites for infectious HIV-1 on platelets and platelet-RBC complexes.

Figure 1. Fractionation of blood from HIV-1-infected patients for obtaining plasma and RBC_native in the presence or absence of EDTA.

Whole blood collected from a patient in the presence (A,B) or absence (C,D) of EDTA, and separated into four plasma and erythrocyte fractions, A, B, C, and D, as shown. Fraction D was treated with 5 mM EDTA, resulting in Fraction E. All of the samples were used immediately for co-culture with PBMC to detect capability for causing HIV-1 infection of the PBMC. Aliquots were stored at −20°C for viral RNA measurements. See Materials and Methods for further details.

Figure 2. PLT-RBC is a subpopulation of cells in whole blood and in preparations of RBC_native from HIV-1-infected patients.

A–B. Flow cytometry was performed with EDTA-anticoagulated whole blood obtained from an HIV-positive individual. A. Representative dot plot shows scatter properties of the RBC, PLT-RBC, and PLT. CD41a+/CD235a+ cells (red) represent PLT-RBC; CD41a+/CD235a- cells (green) represent free PLT; CD41a-/CD235a+ cells (blue) represent RBC; and CD41a-/CD235a- particles (gray) represent small vesicles or debris. B. Quantification of cells shown in frame A that express CD41a and/or CD235a. C–D. Flow cytometry was performed with an EDTA-free RBC_native preparation (see fraction D in Figure 1) obtained from an HIV-positive individual. RBC_native were analyzed as in frames A–B.

Figure 3. Enrichment of PLT-RBC as a subpopulation of normal (uninfected) cells by adding platelets to a preparation of RBC_native from a normal volunteer.

A. Normal RBC_native obtained from citrated blood from a normal (uninfected) volunteer were incubated with platelet-rich plasma, washed, and analyzed by flow cytometry. B. RBC, PLT-RBC, and PLT were separated by sorting the populations shown in frame A. C–E. Light microscopy of Giemsa-stained mixtures of normal RBC_native enriched with PLT-RBC. Magnification: 400X and 1000X, respectively. Arrows in frame C indicate platelets bound to RBC. F. Fluorescent microscopy of PLT-RBC visualized at 630X magnification. Left panel, bright field; 2^nd panel, visualization of both FITC and PE signals; 3^rd panel, visualization of FITC; right panel, visualization of PE signal. Arrows point to platelets.

Figure 4. HIV-1 cell-cell infection of PBMC is dependent on PLT.

A. RBC, PLT-RBC, and PLT were separated by sorting with FACS as shown in Fig. 3B. After incubation with HIV-1_Bal, followed by washing 3 times to remove unbound HIV-1, the cells were co-incubated with PBMC to determine HIV-1 infection. **Cell-cell infection with PLT-bound HIV was higher than with RBC-bound HIV (p = 0.0332) (one-way Anova with Tukey's multiple comparisons test); ***cell-cell infection with PLT-RBC-bound HIV was higher than with RBC-bound HIV (p = 0.0124) (one-way ANOVA with Tukey's multiple comparisons test). B. Mixtures of normal RBC_native enriched with PLT-RBC were prepared by pre-incubation of 0.3 ml containing 10⁹ normal RBC_native with the indicated numbers of PLT in 0.3 ml RPMI dilutions of platelet-rich plasma, and then washed 3 times. The cells were then pre-incubated with HIV-1_Bal, washed 3 times, and examined for HIV-1 infection of PBMC as described in Materials and Methods. Normal RBC_native not pre-incubated with PRP served as a control. *Cell-cell infection was increased as a function of the number of platelets added to RBC_native (p<0.0001, one way Anova).

Figure 5. HIV-1 cell-cell infection of PBMC is eliminated by EDTA.

PLT-RBC enriched RBC_native were prepared by pre-incubation of normal RBC_native (10⁹ cells) with the indicated numbers of PLT, washed 3 times, and pre-incubated with HIV-1_Bal, washed 3 times, and then examined for HIV-1 infection of PBMC in the presence or absence of 5 mM EDTA. Cell-cell infection was eliminated by EDTA (p<0.0001, two-way Anova).

Figure 6. Cell-cell infection of PBMC by RBC_native from HIV-positive patients.

A. EDTA-free RBC_native from 11 chronically-infected patients were obtained as in Fig. 1D and co-incubated with PBMC to examine infection of the PBMC. 25 µl of EDTA-free RBC_native were diluted in 75 µl of IL-2 medium and added to 50 µl of 1.5×10⁵ PHA-stimulated PBMC/well in IL-2 medium. B. 100 µl of EDTA-free plasmas from the patients, obtained as in Fig. 1C and supplemented with 20 U/ml of recombinant IL-2, did not infect co-incubated PBMC. C. The RBC_native shown in frame A were treated with 5 mM EDTA followed by washing 3 times in IL-2 medium before infecting PBMCs (see Fig. 1E) showed no HIV-1 infection of co-incubated PBMC. The infection exhibited by the group of EDTA-free RBC_native (A) was significantly higher using a paired t-test, than the infection exhibited both by the group of EDTA-free plasma (B) and by the group of EDTA-stripped RBC_native (C) at 8 days post-infection (p = 0.0404) and 10 days post-infection (p = 0.038).

Figure 7. Binding of HIV-1 to RBC_native as well as cell-cell infection of PBMC is inhibited by anti-DC-SIGN monoclonal antibody.

A. RBC_native was pre-incubated with HIV-1_Bal in the absence or presence of anti-DC-SIGN mAb in a final 50 µg/ml concentration, washed 3 times, and bound p24 was measured by ELISA. The binding was inhibited by pre-incubation of the cells with the monoclonal antibody (p<0.01, t-test), but not inhibited by pre-incubation of the virus with the monoclonal antibody. B. Cell-cell infection of PBMC was also inhibited in a dose-dependent manner by anti-DC-SIGN mAb (p<0.02 and p<0.01 respectively, paired t-test).