Recently, it has been reported that human ejaculate acts as a potent enhancer of HIV infectivity [5]. This enhancement of infectivity is mediated by a factor, termed Semen-derived Enhancer of Virus Infection (SEVI) [5]. SEVI was identified to be a peptide fragment of the semen marker prostatic acidic phosphatase (PAP) that, upon proteolytic release, forms amyloid fibrils. These fibrils capture HIV virions and direct them to target cells, where they facilitate the fusion of virus and host cell [6].

Interestingly, it has been previously demonstrated that epigallocatechin-3-gallate (EGCG), the major active constituent of green tea, can inhibit the infectivity enhancing effect of SEVI, possibly by interference with de novo SEVI formation or by degradation of present preformed PAP-derived amyloid fibrils [7]. This observation may be important for possible application of EGCG in microbicidal vaginal and rectal gels that could reduce HIV transmission rates [8,9].

Kim et al. recently demonstrated that HIV-1 infectivity enhancement by semen is highly heterogeneous [12]. Here we confirm and extend their finding that the ability of individual semen samples to enhance in vitro HIV-1 infection differs considerably. A minority of samples even inhibited infectivity in our experimental setup. We show that the HIV-1 infectivity enhancing potency is variable even in longitudinal semen samples originating from the same donor obtained at different time points. This finding underscores the high degree of heterogeneity and fluctuations of semen properties, which will pose a significant problem for future interventional trials.

We also used a panel of three different X4- and R5-tropic HIV-1 strains in our experiments, thereby confirming that the infectivity enhancing effect in semen is independent of the viral coreceptor tropism (Figure ​(Figure6B).6B). Synthetic SEVI enhanced HIV-1 infectivity about 25 fold stronger than semen in our experimental setup (Figure ​(Figure5D).5D). This does not contradict our results for semen, as the employed concentration of synthetic SEVI (250 μg/ml) was unphysiologic and much higher than the estimated concentration of SEVI in semen (35 μg/ml, as estimated by Münch et al. [5]). However, these results demonstrate the great potential of synthetic SEVI as an enhancer of in vitro lentiviral infection rates [13].

In contrast to our results, Kim et al. observed overall greater semen-mediated enhancement of HIV-1 infectivity (ranging from 2 to about 50-fold) [12], most likely due to differing experimental conditions and amounts of HIV-virions used for infection. We observed a strong dependence of the absolute HIV-infectivity enhancement on the amount of inoculum (semen-mediated enhancement is approximately threefold stronger when 20 pg BaL-p24 are used for infection instead of the utilized 100 pg). However, the experimental conditions for our study were chosen to ensure maximal test reliability and allow reproducible, consistent results.

In the above mentioned study, Kim et al. have demonstrated that semen-mediated HIV-1 infectivity enhancement correlates with SEVI/PAP248-286 levels measured by ELISA utilizing antisera from guinea pigs and rabbits [12]. Centrifugation through a 100-kDa-pore-size filter removed the entire virus enhancing activity and the reactivity to anti-SEVI antiserum, demonstrating that the enhancing factor has a molecular weight of > 100 kDa, which is in agreement with the fibril hypothesis of infectivity enhancement. A correlation between fibril abundance and HIV-1 infectivity enhancement was also demonstrated by Hauber et al. for synthetic SEVI fibrils, quantified by Congo red staining [7].

Possible causes for the observed heterogeneity of the HIV-infectivity enhancing potency of semen, besides the amount of SEVI fibrils present, could be the variable abundance of cationic polypeptides that inhibit HIV-1 infection [14] and/or naturally occurring proteases in semen. In a recent study, Martellini et al. show that human seminal plasma inhibits fibril formation of PAP248-286 and exhibits proteolytic activity that can inhibit the proviral activity of SEVI under certain conditions [15]. Interestingly, we found a positive correlation between HIV-1 infectivity enhancement and the concentration of zinc in semen by univariate analysis. The findings of Martellini et al. could offer an explanation for this correlation. As Zn²⁺ exhibits inhibitory effects on semen proteases [16], it is intruiging to hypothesize that SEVI fibrils could be more stable in semen samples that contain high amounts of Zn²⁺, and thus tend to have a higher capacity to enhance HIV infectivity.

Our results confirm and extend the original findings by Hauber et al., who first described the inhibition of semen-mediated enhancement of HIV-infectivity by EGCG for a limited number of individual semen samples [7]. While it was thought that the SEVI inhibiting effect by EGCG is mainly mediated through its anti-fibrillogenic properties, we also observed semen-independent inhibition of HIV-infectivity at non-toxic EGCG concentrations. For some samples and for the pool of all samples, the infection rate in presence of EGCG and semen was reduced to below the infection rate of control experiments performed in absence of EGCG and semen. To elucidate this effect, we tested the direct antiviral activity of EGCG in absence of semen in our assay. In contrast to previous results by Hauber et al. with Jurkat 1G5 cells and HIV-1 NL4/3 [7], we observed a semen-independent inhibition of HIV-infectivity by EGCG. The presence of 0.4 mM EGCG in absence of semen resulted in 88.5% inhibition of HIV infectivity. Several other studies also found a direct inhibition of HIV-infectivity by EGCG, albeit in different cells [17-20]. This inhibition is thought to be mediated through direct interaction with the CD4 receptor and/or HIV-1 Env. A direct antiviral activity of EGCG has also been described for other viruses, namely herpes simplex virus (HSV) [21] and hepatitis C virus (HCV) [22]. Ciesek et al. recently demonstrated that EGCG is an inhibitor of hepatitis C virus entry [22]. Regardless of the mechanism, EGCG treatment effectively abolished enhancement of HIV-1 infectivity of the majority of individual semen samples, of pooled semen, and of synthetic SEVI in our experimental setup. Direct inhibition of the infectivity of HIV and other viruses by EGCG would be an additional advantage for its use in antiviral microbicides.

Clinically, the systemic administration of EGCG has been proven to be safe and well tolerated in several studies [23-26]. Topical application of EGCG has been tested in mice with no dermal toxicity detected after ointment application (up to 3% w/w) daily for 30 days [27]. 3% w/w EGCG corresponds to a concentration of 65.5 mM, more than 100-fold higher than the effective concentration used in our assay for inhibition of semen-mediated enhancement of HIV infection. Its safety and low cost together with its effects on other pathogens such as HSV [21] could make EGCG an attractive additional supplement for antimicrobial gels.

In agreement with Münch et al., it has to be pointed out that our results may underestimate the potency of SEVI for in vivo infectivity enhancement [5]. The quantity of HIV-1 virions transmitted during sexual intercourse is considerably lower than the amount of virions used in our assay, and the relative HIV-infectivity enhancement strongly depends on the amount of inoculum. Testing of semen-mediated enhancement of HIV-1 infection in animal models is warranted to clarify its role in sexual transmission of HIV-1.