Human herpesvirus infections can result in a range of diseases, including mononucleosis, genital herpes, encephalitis, various cancers and neurodegenerative diseases. There are nine herpesvirus types that infect humans and five of them are highly prevalent across most human populations: herpes simplex 1 and 2 (HSV-1 and HSV-2), varicella zoster (VZV), Epstein–Barr (EBV) and human cytomegalovirus (HCMV). More than 90% of adults have been infected with at least one of these. An estimated 3.8 billion people under the age of 50 have been infected with HSV-1, the cause of oral herpes; and HCMV, which can lead to hearing loss and developmental delay in newborns, has an estimated seroprevalence of approximately 80%1,2. Despite this, only two approved vaccines exist for only one herpesvirus; both are for VZV, which causes chickenpox and zoster, also known as shingles. So why are there no vaccines available for the other herpesviruses given the substantial morbidity, and in some cases mortality, associated with these viruses?

Herpesviruses have infected and evolved with humans over thousands of years3. During this time, these DNA viruses have adapted to our biology and developed an abundance of mechanisms to counter our immune system. Understanding how our body protects itself from infection and what correlates of protection we should look at during vaccine design remains challenging, since these viruses are not cleared in infected people. After initial acute infection, human herpesviruses persist for life in different niches of the body such as epithelial, endothelial, neuronal or immune cells depending on the virus, with some capable of infecting multiple cell types. Several herpesviruses are also highly species-specific, making animal infection models hard to establish. Finally, herpesviruses can form different viral glycoprotein complexes and utilize a variety of host entry receptors for infection. This results in a broad cell tropism, further complicating the search for appropriate targets of prophylactic interventions.

Understanding which entry complexes exist and how they work mechanistically is a limiting step towards effective therapeutic and vaccine design. Indeed, understanding which host–virus interactions are essential for viral entry can identify targets for antivirals, antibodies or vaccine immunogens to block infection or establish a host immune response. To address this, recent work has uncovered a third HCMV entry complex using cryo-electron microscopy analyses coupled with infection assays4. Norris et al. showed that the complex is abundant on virions and consists of glycoprotein H (gH), which is associated with UL116 and UL141 (an immunoevasin). They name this complex gH-associated tropism and entry-3 (GATE-3) and they recommend renaming other complexes that comprise other viral protein combinations GATE-1 and GATE-2 for ease of reference. Endothelial cells are located on blood vessels and are important for systemic virus spread and persistence. The finding that GATE-3 enhances infection in endothelial cells makes it an interesting candidate for vaccine design in addition to known viral entry components.

Other recent work has uncovered host entry receptors for EBV. Although it was known that EBV could infect epithelial and B cells, the responsible entry receptors were unresolved. In this issue of Nature Microbiology, two studies, one by Wang et al. and another by Zhang et al., as well as a recent study5 published in Nature, show that desmocollin 2 (DSC2), which is expressed on epithelial cells, and R9AP (regulator of G protein signalling 9 binding protein) expressed on epithelial and B cells, are additional host entry receptors for EBV. In a related News & Views article in this issue, Liu and Hahn note that both R9AP and DSC2 interact with the viral gH–gL complex, suggesting that this complex is a potential target for vaccine design.

Alongside knowing which herpesvirus proteins to target for vaccine design, we also need to understand how to elicit neutralizing antibodies using vaccines to then block these interactions. This has been a challenge since most antibody epitopes identified have been non-neutralizing. Two studies, one by Kwong and colleagues in this issue, and another by Grünewald and colleagues6 in Nature, used mutagenesis to stabilize HSV-1 glycoprotein B (gB) in its prefusion form (the virus conformation before it attaches to a host receptor). Prefusion-stabilized viral glycoproteins have previously been shown to elicit neutralizing antibodies that provide protection in respiratory syncytial virus vaccines7. Grünewald and colleagues used HSV-1 prefusion gB to immunize alpacas and identified rare neutralizing nanobodies (antigen-specific single-domain antibody fragments derived from camelid antibodies), whereas Kwong and colleagues immunized mice and also recovered rare antibodies. Both studies used cryo-electron microscopy to resolve the binding modes of these antibodies to gB, which revealed the mechanism by which they block transition from the prefusion to post-fusion form — a conversion needed for virus entry. However, many antibodies specific to the prefusion form did not block this transition. The authors suggest that HSV or herpesviruses in general have evolved mechanisms to protect gB from neutralization by glycan shielding and by developing decoy areas that are similar between pre- and post-fusion states to trigger non-neutralizing responses while protecting sites important for fusion.

Whether neutralization will be the key to effective herpesvirus vaccines or if other immune correlates of protection play a greater role is unclear. A review by Cohen proposes that eliciting virus-specific T cells would be important for developing therapeutic vaccines for EBV, HCMV and HSVs to clear virus infection, as earlier clinical studies showed antiviral effects when giving HCMV- and EBV-specific CD4+ and CD8+ T cells to transplant patients8. Furthermore, different correlates of protection might matter for different patient groups such as pregnant women and immunocompromised transplant patients in the case of HCMV and its respective vaccine design as discussed by Permar and colleagues9. The immune responses triggered by herpesvirus vaccines will likely need to surpass natural immunity, since viruses such as EBV, HSV and HCMV are usually not cleared and are frequently shed in the saliva of infected asymptomatic people. Recent clinical trials for HSV vaccines failed since they did not effectively prevent infection; however, some of the results remain unreported so far and so it is currently unclear where improvements can be made (NCT02837575). Nevertheless, there is still much to learn about immune correlates and causes of protection, which will differ between viruses, sites of infection and patients, to achieve efficacious and maybe even sterilizing vaccines that could prevent transmission of herpesvirus infections.

Whether there is a need to develop vaccines for all human herpesviruses is unclear. Little is known about human herpesviruses 6A, 6B and 7. Nonetheless, vaccinating against herpesviruses can result in additional benefits. For example, individuals that received a recombinant VZV vaccine had a lower risk for developing dementia compared with those receiving influenza and tetanus–diphtheria–pertussis vaccines10. Reports have also linked HSV infection to Alzheimer’s disease, and the persistence of EBV driving multiple sclerosis has been a prominent example of a virus-triggered immune neurodegenerative disease11,12. Whether an EBV vaccine could also prevent multiple sclerosis is an interesting research avenue.

Recent studies have revealed surprising features of herpesviruses that challenge our thinking of traditional vaccine design. This emphasizes the importance of basic research to understand the biology of these viruses, host responses and the long-term consequences of their infections, which ultimately underpins our ability to prevent and treat these infections and tackle this public health need.