Core Framework & Diagram Major Failures in Immunology History
July 24, 20261 Min Read Great Scientists in Immunology History
July 24, 2026Scientific progress has never been a straight line
—— Those costly failures that recalibrated immunology's direction.
I. The STEP trial: the most painful HIV vaccine failure (2007)
The STEP trial (also called the MRKAd5 HIV-1 gag/pol/nef trial) is one of the most destructive failures in HIV vaccine research history — not only because it failed, but because it revealed a risk nobody had anticipated: the vaccine might make some participants more susceptible to HIV infection. This Merck-sponsored trial used adenovirus type 5 (Ad5) as a vector to deliver HIV's gag, pol, and nef gene fragments into the body, theoretically inducing strong CD8+ T cell responses to resist HIV infection. Ad5 is a common human cold virus — many adults had already produced neutralizing antibodies against Ad5 from natural infections (being 'Ad5 seropositive'). In 2007, interim data showed the vaccine provided no protection and, more shockingly, in Ad5 seropositive male participants (those with pre-existing immunity to Ad5), the vaccinated group's HIV infection rate was significantly higher than the placebo group. The clinical trial was immediately halted.
Later analysis suggested this paradox may have come from two mechanisms. First, the Ad5 vector itself activates CD4+ T cells (which are HIV's preferred infection targets) — in individuals who already had Ad5 immunity, this activation was stronger, recruiting large numbers of CD4+ T cells to the vaccine injection site, actually providing more infection targets for HIV. Second, 'original antigenic sin' — existing Ad5 immune memory may have interfered with the vaccine's response to HIV antigens, making immune memory primarily respond to the adenovirus vector rather than HIV antigens. The STEP trial failure drove deep reflection in the HIV vaccine field: adenovirus vector selection must consider pre-existing immunity in populations; the type of vaccine-induced immune activation must consider not just 'how many T cells were activated' but 'what type of T cells, activated where.' This lesson later influenced COVID vaccine design — adenovirus vector COVID vaccines all considered this issue to some degree.
2. Sepsis anti-inflammatory drug serial failures: the animal model trap
From the 1990s to 2000s, over thirty anti-inflammatory drugs targeting sepsis (the systemic infection-triggered excessive inflammatory response) failed consecutively in large-scale Phase III clinical trials — despite many showing convincing effects in animal experiments. The scale of this failure was nearly unmatched in clinical medicine history. Sepsis animal models typically use cecal ligation and puncture (CLP) or intraperitoneal lipopolysaccharide (LPS) injection to simulate infection-induced systemic inflammation. In these models, early administration of TNF-α inhibitors, IL-1 receptor antagonists, PAF antagonists, and other anti-inflammatory drugs could significantly reduce animal mortality, sometimes quite dramatically. But when these drugs entered human trials, they repeatedly proved ineffective, in some cases harmful. After years of failures, researchers gradually recognized several fundamental differences. First, timing problem: mouse sepsis models required giving anti-inflammatory drugs immediately after infection (sometimes even prophylactically before infection) to be effective; clinical patients typically present hours to days after infection starts, with inflammation already in a different stage. Second, pathological heterogeneity: human sepsis is an extremely heterogeneous disease — different pathogens, different infection sites, different underlying diseases result in enormously different immune pathological mechanisms; treating all sepsis patients with a single drug was destined to have limited effectiveness. Third, early pro-inflammatory vs. late immunosuppression: sepsis courses divide into early (highly pro-inflammatory) and late (immunosuppression, immune function exhaustion) stages, while many clinical trials enrolled patients from both stages, diluting potential efficacy signals.
3. High-dose IL-2's dark chapter: cancer immunotherapy's price
Before immune checkpoint inhibitors appeared, cancer immunology experienced a long 'dark age' — between the 1970s and 1990s, various immune activation attempts repeatedly failed, making 'cancer immunotherapy' almost a scientific laughingstock. High-dose interleukin-2 (HD IL-2)'s clinical experience was this period's epitome. IL-2 is T cell proliferation's core signal molecule. Theoretically, administering exogenous IL-2 should powerfully activate anti-tumor T cells. Steven Rosenberg's team at NIH systematically explored HD IL-2 for advanced melanoma and renal cell carcinoma from the 1980s. In some patients, astonishing complete remissions appeared — some patients' tumors completely disappeared, with remissions lasting years to over a decade. But HD IL-2's side effects were extremely severe: it induces capillary leak syndrome, causing massive intravascular fluid to leak into tissue spaces, resulting in hypotension, pulmonary edema, and kidney dysfunction. Patients receiving HD IL-2 almost universally required intensive care; treatment-related mortality was initially four to five percent, improving to one to two percent with systematic supportive care. Such severe toxicity meant HD IL-2 could only be administered at a small number of large medical centers with specialized critical care capacity, and only relatively young, physically robust patients could tolerate it. HD IL-2's lesson profoundly influenced later immunotherapy design philosophy: systemic, non-specific immune activation is dangerous; precise, local, targeted-at-specific-checkpoint adjustment is sustainable treatment strategy. This directly provided a negative control for PD-1/CTLA-4 checkpoint inhibitor design: checkpoint inhibitors' advantage is precisely that they precisely release specific brakes, not broadly floor the accelerator.
4. HIV vaccines: thirty years of continuing failure
HIV vaccines are medical history's longest-running vaccine development failure. Since US Secretary of Health Margaret Heckler declared in 1984 that 'an HIV vaccine will be developed within two years,' billions of dollars of research investment and dozens of clinical trials have yet to produce a single approved effective vaccine. This isn't one failure but a series of failures in different directions, each pushing the field one step forward but seemingly always distant from the finish line. The main obstacles: HIV's high mutation rate — the virus's envelope protein (gp120/gp41) mutates rapidly; any vaccine designed based on a specific variant quickly faces a virus that has undergone antigenic drift. Additionally, HIV evolved sophisticated structures shielding neutralization epitopes — key sites recognizable by neutralizing antibodies are often covered by glycosylated 'glycan shields,' or only exposed during brief transitional states when the virus is fusing with cells. Researchers spent twenty years beginning to understand how to induce 'broadly neutralizing antibodies' (bNAb), which can cover HIV's conserved epitopes — but even understanding the principle, how to effectively induce them through vaccines remains unsolved. HIV vaccine failures drove deep progress across multiple immunology and vaccinology directions: broadly neutralizing antibody discovery and characterization provided a research framework for broad-spectrum influenza vaccines; deep research into immune memory provided theoretical foundations for mRNA vaccine technology; and research on HIV immune escape mechanisms directly contributed to universal understanding of relationships between viral antigen variation and immune pressure. Sometimes, the longer a failure persists, the deeper the scientific accumulation it drives.
5. Methodological revolution extracted from failures
These major failures directly drove the evolution of modern clinical research methodology. Stricter preclinical-to-clinical translation standards: sepsis trials' serial failures drove requirements for animal models closer to human disease before clinical trials, stricter statistical analysis (independent validation, not just single laboratory replication), and more systematic biomarker research (determining which patient types might benefit before designing clinical trials). Adaptive clinical trial design: traditional fixed-design clinical trials (predetermined number, endpoints, duration; analyze after trial ends) are inefficient and difficult to adjust based on interim data. HIV vaccine and sepsis trial failures drove broad adoption of 'adaptive clinical trial' design — allowing adjustment of sample size, inclusion criteria, or comparison groups based on interim data during the trial, improving efficiency of discovering real signals. Patient stratification and precision medicine thinking: sepsis and early cancer immunotherapy failures largely resulted from treating highly heterogeneous patient groups as one. This drove biomarker-driven patient stratification — using genetic, protein, or immune phenotype to predict which type of patients will respond to specific treatments, testing the right drug in the right patients. Today's tumor immunotherapy patient selection (based on PD-L1 expression, TMB, MSI) embodies this methodology.
Frequently Asked Questions