Platelets, anucleated fragments of megakaryocytes, have a central role in primary and secondary haemostasis. It is becoming appreciated, however, that platelets have functions beyond ensuring the integrity of the vasculature, which include a role in defence against invading pathogens [7, 28]. We found that activation of platelets induces release of one or more antiviral factors, which suppress HIV-1 infection at the stage of viral entry. Inhibition of several HIV-1 strains was observed and the antiviral activity was independent of coreceptor tropism or target cell type. Recombinant CXCL4 inhibited HIV-1 infection and the HIV-1 inhibition by platelet supernatants was largely rescued by a neutralizing CXCL4-specific antibody, indicating a major contribution of CXCL4 to HIV-1 inhibition by platelets. These results suggest that platelets could constitute a so far unappreciated innate defence against HIV-1 infection.
It is well documented that platelets associate with HIV-1 in cell culture and in infected patients [11–13, 29, 30]. Our previous results  and a study by Boukour and colleagues  demonstrated that platelets bind HIV-1 mainly via the C-type lectin DC-SIGN and that bound virions are infectious for adjacent target cells . The C-type lectin-like protein CLEC-2 also contributed to HIV-1 capture by platelets but capture efficiency was reduced compared to DC-SIGN [15, 31]. Our present study adds MR to the list of C-type lectin (−like) proteins which could contribute to HIV-1 capture by platelets and potentially megakaryocytes. Although a contribution of MR to HIV-1 interactions with platelets remains to be demonstrated, these findings highlight that platelets express several lectins involved in pathogen recognition and might thus modulate pathogen spread and pathogen specific immune responses.
The presence of platelets in HIV-1 infected T cell cultures reduced viral spread efficiently and in a dose-dependent manner. At first sight, this finding is counterintuitive, considering our previous finding that platelets capture and transfer infectious HIV-1 to T cells via lectins . However, different experimental conditions were chosen to generate these results, and the platelet/T cell co-cultures infected with low amounts of HIV-1 in the absence of a wash step (Figure 2A, B of present manuscript) reflect the physiological situation better than the conditions previously chosen to analyse viral capture (high amount of input virus, removal of unbound virus from platelets ). The inhibition of HIV-1 observed in these co-culture experiments was maximal when platelet – T cell ratios were used that were similar to those found in human blood, suggesting that platelet-dependent blockade of HIV-1 might occur in patients.
Platelet granules contain hundreds of bioactive molecules which are released into the extracellular space upon platelet activation . Two findings indicate that platelet granules contain an anti-HIV-1 activity, which is released upon activation, and which is largely responsible for the inhibition of HIV-1 spread in platelet/T cells co-cultures: Activated and subsequently washed (and thus granule-deprived) platelets inhibited HIV-1 spread less efficiently than untreated platelets and supernatants of activated but not resting platelets efficiently suppressed HIV-1 infection. Such a scenario raises two immediate questions. How are platelets activated in HIV-1 infected co-cultures with T cells and what is the nature of the antiviral factor? Concerning the trigger for activation, direct contact between HIV-1 and platelets could be sufficient for activation, as previously demonstrated for certain adenoviruses . However, exposure of platelets to virus-like particles bearing HIV-1 Env did not induce appreciable platelet activation (Additional file 4: Figure S4; Additional file 5: Additional methods), arguing against this hypothesis. Alternatively, activation of platelets might be relatively unspecific and could be induced by contact of platelets with T cells or tissue culture plastic during culture – a scenario that we currently favour (Figure 1A and Additional file 6: Figure S5).
Our initial attempts to identify the antiviral factor(s) released by platelets focussed on ligands for the HIV-1 coreceptors CCR5 and CXCR4, the CC-chemokines CCL3 (MIP-1α), CCL4 (MIP-1β), CCL5 (RANTES) and the CXC-chemokine CXCL12 (SDF1), respectively. These chemokines can block HIV-1 entry and release of CXCL12 was detected upon platelet activation (data not shown), in agreement with a previous study . However, the concentration of CXCL12 was below that previously shown to block cellular entry of X4-tropic HIV-1. In addition, neutralizing antibodies directed against CXCL12 did not rescue HIV-1 inhibition and no evidence for downregulation of CXCR4 by platelet supernatants was obtained (data not shown), indicating that platelet-derived CXCL12 was not involved in blockade of X4-tropic HIV-1. A report published during the preparation of this manuscript showed that the chemokine CXCL4 binds to HIV-1 Env and inhibits HIV-1 entry . In contrast, cellular entry of HIV-2 and SIV was not blocked by CXCL4 . CXCL4 is present at high levels in α-granules of platelets [21, 22] and we therefore focussed our further analysis on this chemokine. Platelet supernatants and recombinant CXCL4 exerted antiviral activity against HIV-1 without interfering with CD4 and coreceptor expression (Additional file 7: Figure S6) and neutralization of CXCL4 largely prevented HIV-1 inhibition by platelet supernatants, demonstrating that inhibition of HIV-1 infection by activated platelets is mainly due to release of CXCL4. Recombinant CXCL4 and endogenous CXCL4 in platelet supernatants also inhibited MLV Env-driven host cell entry. Whether the blockade of MLV Env by CXCL4 also involves interactions between these proteins or is due to CXCL4 binding to host cell factors required for MLV Env-dependent entry remains to be determined.
Release of CXCL4 by platelets could impact HIV-1 dissemination between individuals. Thus, transmission of HIV-1 via the sexual and particularly the parenteral route frequently involves (micro-) vascular injury, and the resulting platelet activation and release of CXCL4 might reduce transmission efficiency. Platelet-derived CXCL4 might also modulate viral spread during the chronic phase of the infection: Platelets isolated from the blood of HIV-1 patients were reported by several [18, 19] but not all  studies to express activation markers. Moreover, a recent analysis demonstrated that platelets from HIV-1 infected individuals have a reduced threshold to activation, and that plasma from HIV-1 patients activates platelets obtained from healthy donors . Thus, the activation status of platelets is increased in the context of HIV-1 infection, potentially due to pro-inflammatory cytokines or invading bacteria , which are present at elevated levels in HIV-1 patients  and are known to activate platelets [36, 37]. As a consequence, platelets might constantly release CXCL4, which would explain why viral load and platelet counts were found to be inversely correlated in infected humans  and why a direct correlation between CD62P levels and viral load was observed in a recent study . However, it also needs to be noted that CXCL4 can increase HIV-1 replication in macrophages after successful viral entry into these cells , suggesting that CXCL4 might impact viral spread via more than one mechanism.
Collectively, platelets can negatively regulate HIV-1 spread in an activation status-dependent manner by release of CXCL4 and might form an innate defence against HIV-1. These findings are in line with the observations that platelets express immune cell lectins and toll-like receptors  and employ these receptors to respond to pathogen invasion . Ultimately, proof for a protective role of platelets in HIV-1 infection must come from animal models or patients with specific defects in platelet function.