Enfuvirtide: The HIV Peptide That Blocks Viral Fusion

Antiviral Peptide Drugs

1.96 log₁₀ viral load drop

In the first human proof-of-concept trial, enfuvirtide at 100 mg twice daily reduced plasma HIV RNA by a median of 1.96 log₁₀ copies/mL in 14 days, comparable to triple-drug antiretroviral regimens.

Kilby et al., Nature Medicine, 1998

Kilby et al., Nature Medicine, 1998

View as image

In 1994, a team at Duke University synthesized a 36-amino-acid peptide called DP-178 that blocked HIV-1 infection at concentrations below 5 ng/mL. That peptide became enfuvirtide, sold as Fuzeon, and in March 2003 it became the first HIV fusion inhibitor approved by the FDA and the first antiretroviral drug that worked entirely outside the host cell.^[1] Its mechanism was unlike anything in the existing HIV drug arsenal: rather than targeting viral enzymes after the virus had already entered a cell, enfuvirtide physically prevented the virus from fusing with the cell membrane in the first place. The drug demonstrated that peptides could serve as frontline therapeutics against a virus that had resisted every other approach. It also exposed the limits of peptide drugs: twice-daily subcutaneous injections, injection site reactions in 98% of patients, over 100 manufacturing steps, and a price tag exceeding $25,000 per year. Enfuvirtide was discontinued from the U.S. market in February 2025, replaced by newer oral antiretrovirals that are easier to take and cheaper to produce. But its legacy extends far beyond HIV treatment. Enfuvirtide proved that targeting viral entry with a designed peptide was viable in humans, a concept that has since driven development of peptide-based entry inhibitors for hepatitis B and D, influenza, and coronaviruses. This article covers the full arc: from DP-178 in a Duke laboratory to FDA approval, the clinical trial evidence, the manufacturing challenges that constrained its use, resistance patterns, and the next-generation peptide fusion inhibitors it inspired. For the complete list of approved antiviral peptides, see Every FDA-Approved Antiviral Peptide. For the related hepatitis C peptide drugs, see Boceprevir and Telaprevir.

How HIV enters a cell and where enfuvirtide intervenes

HIV-1 entry is a multi-step process that begins when the viral surface glycoprotein gp120 binds to the CD4 receptor on the target cell. This initial attachment triggers a conformational change in gp120 that exposes a binding site for a chemokine co-receptor, either CCR5 or CXCR4. Once gp120 engages both receptors, the transmembrane glycoprotein gp41 undergoes a dramatic structural rearrangement. The N-terminal heptad repeat region (HR1) of gp41 forms a trimeric coiled-coil, and the C-terminal heptad repeat region (HR2) folds back against it, creating a six-helix bundle that pulls the viral and cell membranes together, forcing fusion.^[6]

Enfuvirtide is a 36-amino-acid peptide that mimics the HR2 region of gp41. By binding to the HR1 trimer during the transient window between co-receptor engagement and six-helix bundle formation, enfuvirtide prevents HR2 from folding back. Without six-helix bundle formation, the membranes cannot merge. The virus is left stranded outside the cell, unable to deliver its genetic payload.^[1]

This mechanism is fundamentally different from all other classes of antiretroviral drugs. Nucleoside reverse transcriptase inhibitors (NRTIs), non-nucleoside reverse transcriptase inhibitors (NNRTIs), protease inhibitors, and integrase inhibitors all target viral processes that occur inside the cell after infection has already happened. Enfuvirtide acts before infection begins, in the extracellular space, at the moment of membrane fusion. This extracellular mechanism means the drug does not need to enter the cell, does not compete with intracellular drug transporters, and has minimal interactions with other antiretrovirals.^[3]

The discovery: from DP-178 to T-20

Enfuvirtide originated from a systematic peptide mapping study at Duke University. Wild et al. (1994) synthesized a series of peptides corresponding to different regions of the HIV-1 gp41 transmembrane protein and tested their ability to inhibit viral infection. One peptide, designated DP-178, corresponding to residues 643-678 of the HIV-1LAI isolate, stood out with extraordinary potency.^[1]

DP-178 blocked 100% of virus-mediated cell-cell fusion at concentrations below 5 ng/mL, with an IC90 of approximately 1.5 ng/mL. It reduced infectious titer of cell-free virus at roughly 80 ng/mL. The therapeutic window was wide: cytotoxicity and cytostasis were detected only at peptide concentrations 10,000 to 100,000 times higher than the inhibitory dose. PCR analysis of newly synthesized proviral DNA confirmed that DP-178 blocked an early step in the viral life cycle, prior to reverse transcription, consistent with entry inhibition rather than post-entry interference.^[1]

The inhibitory activity was HIV-1 specific. Blocking an HIV-2 isolate required 100 to 1,000 times more peptide, reflecting sequence divergence in the gp41 regions between HIV-1 and HIV-2. DP-178 was active against both laboratory-adapted (prototypic) and primary clinical HIV-1 isolates, an important finding because primary isolates often resist compounds that work only on lab strains.^[1]

The researchers at Duke formed Trimeris, Inc. to develop DP-178 clinically. The peptide was renamed T-20, and in 1999 Trimeris partnered with Hoffmann-La Roche to fund the large-scale clinical trials and solve the manufacturing challenges that a 36-amino-acid therapeutic peptide would inevitably present.^[3]

First proof in humans: the 1998 Nature Medicine study

The transition from in vitro to in vivo proof-of-concept came in 1998. Kilby et al. administered intravenous T-20 as monotherapy for 14 days to sixteen HIV-infected adults in four dose groups: 3, 10, 30, and 100 mg twice daily.^[2]

The results were dose-dependent and striking. All four subjects in the highest dose group (100 mg twice daily) achieved plasma HIV RNA below 500 copies/mL by branched-chain DNA (bDNA) assay. A more sensitive RT-PCR assay (detection threshold 40 copies/mL) showed that while undetectable levels were not achieved in the 14-day dosing period, there was a median decline of 1.96 log₁₀ copies/mL in plasma HIV RNA in these subjects. Significant, dose-related viral load declines occurred in all subjects receiving 10 mg or higher.^[2]

Short-term administration appeared safe. No dose-limiting toxicities were observed. The study provided proof-of-concept that viral entry could be successfully blocked in vivo, and that the magnitude of viral suppression was comparable to existing triple-drug antiretroviral regimens. This was a pivotal moment for the field: it demonstrated that a designed peptide, administered systemically, could suppress a chronic viral infection in humans by preventing cellular entry.^[2]

Phase 3 trials: TORO-1 and TORO-2

The pivotal Phase 3 studies that led to FDA approval were TORO-1 (T-20 vs. Optimized Regimen Only, Study 1) and TORO-2, both published in the New England Journal of Medicine in 2003.^[3]

Both trials enrolled heavily treatment-experienced patients, individuals who had received and failed drugs from all three existing antiretroviral classes (NRTIs, NNRTIs, and protease inhibitors) and had plasma HIV-1 RNA above 5,000 copies/mL. Patients were randomized to receive either 90 mg enfuvirtide subcutaneously twice daily plus an optimized background regimen (OBR), or OBR alone.

TORO-1 (North and South America, Lalezari et al.): At 24 weeks, the mean viral load reduction was 1.696 log₁₀ copies/mL in the enfuvirtide group versus 0.764 log₁₀ in the control group (P<0.001). Mean CD4+ cell count increases were 76 cells/mm³ versus 32 cells/mm³ (P<0.001).^[3]

TORO-2 (Europe and Australia, Lazzarin et al.): At 24 weeks, the mean viral load reduction was 1.429 log₁₀ in the enfuvirtide group versus 0.648 log₁₀ in controls. CD4+ increases were 65.5 cells/mm³ versus 38.0 cells/mm³.^[3]

Combined 48-week data showed that 30% of patients receiving enfuvirtide plus OBR achieved HIV-1 RNA below 400 copies/mL, compared to 12% with OBR alone. The mean CD4+ increase at 48 weeks was 91 cells/mm³ versus 45 cells/mm³.^[3]

These trials enrolled the most difficult-to-treat HIV population: patients who had exhausted available options. The fact that enfuvirtide could achieve virologic suppression in a meaningful proportion of these patients validated the fusion inhibitor class and earned the drug accelerated FDA approval on March 13, 2003.

Injection site reactions and the practical burden

Enfuvirtide's safety profile was dominated by one issue: injection site reactions (ISRs). In the TORO trials, 98% of patients receiving enfuvirtide experienced ISRs, including pain, induration (hardening), erythema, nodules, and cysts. Most were mild to moderate and did not require analgesics or limit daily activities, but the reactions were essentially universal.^[3]

The injection burden went beyond discomfort. Enfuvirtide required twice-daily subcutaneous injections, each involving reconstitution of a lyophilized powder with sterile water, followed by a waiting period for complete dissolution. For patients already taking multiple oral antiretrovirals, adding a complex injection regimen with near-certain injection site reactions created an adherence challenge that limited the drug's practical utility.

Other adverse events included bacterial pneumonia (occurring at a higher rate in the enfuvirtide group than in controls), peripheral neuropathy, insomnia, and eosinophilia. No drug-drug interactions with other antiretrovirals were identified, which was expected given enfuvirtide's extracellular mechanism and peptide-based metabolism.^[3]

The manufacturing challenge: 100 steps to build a peptide drug

Enfuvirtide was the most complex antiretroviral drug ever manufactured, and its production costs shaped the drug's entire commercial trajectory. A standard small-molecule antiretroviral involves 8 to 10 production steps. Enfuvirtide required over 100.^[3]

The manufacturing process used convergent solid-phase peptide synthesis (SPPS). Three separate peptide fragments were assembled individually, each built by sequential coupling of amino acids onto a solid resin support. The three fragments were then joined to create the full 36-amino-acid chain, which was cleaved from the resin, purified, and formulated. The process required 45 kg of raw materials to produce 1 kg of finished drug. Roche built a dedicated manufacturing facility in Boulder, Colorado, at a cost exceeding $100 million.

The resulting price was approximately $25,000 per patient per year, making enfuvirtide one of the most expensive antiretrovirals on the market. Production capacity was initially constrained to supply only 12,000 patients in the first year and 32,000 in the second year. These constraints meant that even patients who could benefit from enfuvirtide might not be able to access it.

The manufacturing challenge was not unique to enfuvirtide. It exemplified a fundamental limitation of peptide therapeutics: solid-phase synthesis generates substantial chemical waste, costs increase sharply with peptide length, and scaling production to meet chronic disease treatment demands is far more difficult than scaling small-molecule drug production.^[7] The enfuvirtide manufacturing experience directly influenced the peptide drug industry's subsequent investment in recombinant production methods, continuous-flow synthesis, and chemical ligation strategies designed to reduce cost and waste for long peptide sequences.

Resistance: how HIV evades a fusion inhibitor

HIV-1 develops resistance to enfuvirtide through mutations in the HR1 region of gp41, the same region the drug targets. The primary resistance mutations cluster in a 10-amino-acid stretch at positions 36-45, with changes at codons 36, 38, 42, 43, and 44 being the most common. The V38A substitution appears most frequently, followed by G36D/E, N42D, and V38M. These mutations confer 15- to 445-fold reductions in enfuvirtide susceptibility.^[6]

Resistance emerges relatively quickly under enfuvirtide monotherapy or when the optimized background regimen provides inadequate viral suppression, allowing viral replication to continue in the presence of drug selection pressure. In the TORO trials, resistance mutations were detected in the majority of patients who experienced virologic failure on enfuvirtide-containing regimens.

One unexpected finding was that certain enfuvirtide resistance mutations, particularly V38A and V38E, were associated with sustained CD4+ cell count increases even in patients with detectable viral loads. This paradox suggested that the resistance mutations imposed a fitness cost on the virus, reducing its replicative capacity or pathogenicity even as they conferred drug resistance. The clinical implication was that continuing enfuvirtide in the presence of genotypic resistance might still provide immunologic benefit.^[6]

Cross-resistance to other fusion inhibitor peptides depends on the specific mutations selected and the binding footprint of the alternative peptide. Newer peptide fusion inhibitors have been designed specifically to maintain activity against enfuvirtide-resistant variants by targeting different or extended regions of the gp41 coiled-coil.^[6]

The structural biology of enfuvirtide binding

Understanding why enfuvirtide works, and why resistance mutations can disrupt it, required detailed structural studies of the peptide-protein interaction. Enfuvirtide adopts an alpha-helical conformation when it binds to the HR1 trimeric coiled-coil of gp41. The alpha-helix is not inherently stable in the unbound peptide; it forms upon contact with the target, a process called induced folding.

Stocks et al. (2021) used ion mobility mass spectrometry (IM-MS) and hydrogen-deuterium exchange (HDX) to characterize discrete conformational states of native enfuvirtide and hydrocarbon-stapled variants. The study found that native enfuvirtide exists as a conformational ensemble in solution, with only a fraction of molecules adopting the alpha-helical structure needed for target binding. Introducing hydrocarbon staples at specific positions dramatically shifted the conformational equilibrium, increasing alpha-helical content and reducing the structural flexibility of the peptide.^[4]

The location of the staple mattered. Staples placed at positions that reinforced the binding face of the helix increased target affinity. Staples at other positions constrained the peptide in non-productive conformations. This study provided a structural rationale for why enfuvirtide derivatives with enhanced helical stability might overcome some of the parent drug's limitations, including susceptibility to proteolytic degradation and the need for injection.^[4]

What enfuvirtide taught the peptide drug field

Enfuvirtide was a proof-of-concept that rippled across virology, pharmacology, and peptide chemistry. Cooper and Bhatt (2004) reviewed its significance in The Lancet Infectious Diseases, noting that enfuvirtide established several principles that have guided peptide antiviral development since.^[3]

First, targeting viral entry is therapeutically viable. Before enfuvirtide, all approved antiretrovirals acted after the virus had already entered the cell. The success of enfuvirtide opened the entry step as a druggable target, a concept subsequently exploited by bulevirtide for hepatitis D, the CCR5 antagonist maraviroc (a small molecule inspired by the same entry-blocking concept), and experimental peptide fusion inhibitors against influenza and coronaviruses.^[6]

Second, peptides can achieve pharmacologically relevant concentrations in human plasma and maintain antiviral pressure over clinically meaningful timeframes, even against a fast-mutating virus like HIV-1.^[2]

Third, the practical limitations of enfuvirtide, injection-only administration, injection site reactions, manufacturing complexity, and high cost, defined the engineering challenges that the next generation of peptide drugs would need to solve. These were not reasons to abandon peptide therapeutics, but rather a to-do list for the field.

Tsomaia (2015) placed enfuvirtide in the broader landscape of peptide therapeutics, noting that over 60 peptide drugs had reached the market and more than 140 were in clinical trials, with enfuvirtide among the most instructive examples of both the potential and the constraints of the platform.^[7]

Next-generation peptide fusion inhibitors

Enfuvirtide's limitations drove a wave of research into improved peptide fusion inhibitors. The goal: retain the mechanism while solving the problems of proteolytic instability, injection-only delivery, and narrow resistance profiles.

Hydrocarbon-stapled peptides represent one approach. By introducing covalent carbon-carbon bridges across one or two helical turns of the peptide backbone, chemists can lock the peptide in its bioactive alpha-helical conformation, increase resistance to proteolytic digestion, and improve membrane permeability. Stocks et al. (2021) demonstrated that stapling enfuvirtide altered its conformational ensemble in measurable ways, with staple placement determining whether the modification enhanced or impaired function.^[4]

Double-stapled short peptides push the concept further. Wang et al. (2024) described D26, a hydrocarbon double-stapled helical peptide that inhibits HIV-1 infection and directly inactivates cell-free virions. Unlike enfuvirtide, which acts passively by blocking fusion, D26 functions as a virus inactivator that can attack and disable HIV-1 particles without relying on the viral replication cycle. The double stapling provided high protease resistance, an extended in vivo half-life, enhanced penetration into tissue sanctuary sites where HIV persists, and detectable oral bioavailability.^[5] An orally available peptide fusion inhibitor would eliminate the injection burden that limited enfuvirtide's clinical adoption.

Peptide-liposome conjugates take a different approach to the delivery problem. Erdmann et al. (2025) reported on an HIV-1 gp41 peptide-liposome vaccine in the HVTN133 Phase 1 trial, using peptides from the gp41 membrane-proximal external region (MPER) conjugated to liposomes to elicit broadly neutralizing antibody precursors. While this is a vaccine rather than a therapeutic, it builds on enfuvirtide's fundamental insight: the gp41 fusion machinery is a vulnerable and targetable structure on the virus surface.^[9]

AI-designed antiviral peptides represent the newest frontier. Mashhadi et al. (2025) reviewed how artificial intelligence is being applied to antiviral peptide design, including deep learning models that predict peptide-target interactions, generative models that create novel sequences with desired properties, and delivery platform engineering that addresses the pharmacokinetic challenges enfuvirtide faced. The review notes that enfuvirtide's approval catalyzed the field of antiviral peptide therapeutics and that AI-driven design methods now offer solutions to the structure-activity and delivery challenges that constrained the first generation.^[8]

Pu et al. (2019) comprehensively cataloged protein- and peptide-based HIV entry inhibitors developed after enfuvirtide, including those targeting gp120 (the attachment step) rather than gp41 (the fusion step). The review documented over a dozen peptide candidates in preclinical or early clinical development, many designed specifically to overcome enfuvirtide resistance by binding different epitopes on the fusion machinery or by incorporating structural modifications that improve stability.^[6]

Why enfuvirtide was discontinued

Genentech ceased marketing and commercial distribution of Fuzeon in the United States as of February 28, 2025. The decision reflected several converging factors rather than any safety concern.

The HIV treatment landscape had changed fundamentally since 2003. Newer antiretroviral classes, including integrase strand transfer inhibitors (dolutegravir, bictegravir, cabotegravir) and maturation inhibitors, offered potent viral suppression with oral dosing, fewer side effects, and lower costs. Long-acting injectable formulations of cabotegravir plus rilpivirine provided monthly or bimonthly injections that replaced daily oral regimens, addressing adherence concerns without the injection site reaction burden of enfuvirtide.

The treatment-experienced population that enfuvirtide was designed for had also shrunk. The success of combination antiretroviral therapy and the availability of potent regimens with high barriers to resistance meant fewer patients exhausted all other options. The number of patients who needed enfuvirtide as a component of a salvage regimen declined year over year.

Manufacturing economics also played a role. Maintaining a dedicated production facility for a 36-amino-acid peptide drug serving a diminishing patient population was commercially unsustainable. The high per-patient cost that reflected genuine manufacturing complexity, not pricing strategy, made enfuvirtide difficult to justify when alternatives existed.

In the UK, Fuzeon injection was similarly discontinued in January 2025.

The evidence landscape: what holds up, what remains uncertain

The enfuvirtide evidence base is mature. The discovery studies (Wild et al., 1994; Kilby et al., 1998) and Phase 3 trials (TORO-1 and TORO-2, 2003) were conducted with rigorous methodology and published in high-impact journals. The drug's mechanism of action is well-characterized at the molecular level, and its clinical efficacy in treatment-experienced patients was definitively established.

Several limitations should be noted. Enfuvirtide was never tested in treatment-naive patients in a Phase 3 trial, because its injection-based delivery made it inappropriate as first-line therapy. There are no long-term (multi-year) randomized trial data because the drug was always used as part of a salvage regimen, and patients were typically transitioned to other agents when they became available. Resistance emerged in a substantial proportion of patients who did not achieve virologic suppression, though the fitness cost of resistance mutations partly mitigated the clinical impact.

The manufacturing and cost challenges were real constraints on access, not artifacts of pricing decisions. The 36-amino-acid length placed enfuvirtide at the upper end of what solid-phase synthesis could economically support, and the experience directly shaped subsequent investment in alternative production methods for therapeutic peptides.^[10]

The Bottom Line

Enfuvirtide was the first peptide-based HIV drug and the first antiretroviral to block viral entry by preventing membrane fusion. Clinical trials demonstrated clear efficacy in treatment-experienced patients, with viral load reductions of 1.7 log₁₀ at 24 weeks in the TORO-1 trial. Its legacy is primarily conceptual: enfuvirtide proved that peptide-based entry inhibitors could work in humans and defined the engineering challenges (stability, delivery, cost) that next-generation antiviral peptides are now solving with stapling chemistry, AI-driven design, and improved manufacturing methods.

Frequently Asked Questions

What is enfuvirtide and how does it work?

Enfuvirtide is a 36-amino-acid synthetic peptide that blocks HIV-1 from entering human cells. It binds to the HR1 region of the gp41 transmembrane protein during the fusion process, preventing the six-helix bundle formation required for viral and cell membrane fusion. Wild et al. (1994) first demonstrated this peptide blocked 100% of HIV-1 cell-cell fusion at concentrations below 5 ng/mL.

Why was Fuzeon discontinued?

Genentech stopped marketing Fuzeon in the United States in February 2025. The decision reflected the availability of newer, more convenient antiretrovirals (oral integrase inhibitors, long-acting injectables), a shrinking population of treatment-experienced patients who needed salvage therapy, and the commercially unsustainable cost of manufacturing a 36-amino-acid peptide requiring over 100 production steps.

How effective was enfuvirtide in clinical trials?

In the Phase 3 TORO-1 trial, enfuvirtide plus an optimized background regimen reduced HIV-1 viral load by 1.696 log₁₀ copies/mL at 24 weeks, compared to 0.764 log₁₀ for background regimen alone. At 48 weeks, 30% of patients in the enfuvirtide arm achieved viral loads below 400 copies/mL versus 12% in the control arm.

What were the main side effects of enfuvirtide?

Injection site reactions occurred in 98% of patients, including pain, hardening, redness, and nodules. These were typically mild to moderate. Other reported adverse events included an increased rate of bacterial pneumonia, peripheral neuropathy, and insomnia. No significant drug-drug interactions with other antiretrovirals were identified due to enfuvirtide's extracellular mechanism.

Can HIV become resistant to enfuvirtide?

HIV-1 develops resistance through mutations in the HR1 region of gp41, primarily at positions 36-45. The V38A substitution is most common, with resistance mutations conferring 15- to 445-fold reductions in drug susceptibility. Resistance emerged in the majority of patients who experienced virologic failure on enfuvirtide, though some resistance mutations imposed a fitness cost that preserved partial immunologic benefit.

Are there newer peptide fusion inhibitors being developed?

Multiple next-generation approaches are in development. Hydrocarbon-stapled enfuvirtide variants show increased alpha-helical stability and protease resistance. Wang et al. (2024) reported a double-stapled short peptide (D26) with oral bioavailability and direct virion-inactivating activity. AI-driven design methods are also being applied to create novel antiviral peptide sequences with improved pharmacokinetic properties.