Video summary

약을 이해하려면 단백질을 알아야 한다? 원리부터 알려드림 (feat. 김성훈 교수) [취미는 과학/ 60화 확장판]

Main summary

Key takeaways

Educational

Main ideas & concepts (lesson-style summary)

Greatest medicines and why they matter

The discussion compares historically major drugs/approaches:

  • Ethanol (“water of life”) as an early life-saving infection treatment.
  • Aspirin as a widely used anti-inflammatory painkiller for minor pain and inflammation.
  • Antibiotics as arguably the medicine that saved the most lives, dramatically changing mortality trends once developed (especially alongside vaccines).

A humorous counterpoint frames a mother’s hand causing immediate pain relief as a form of placebo, highlighting that “medicine-like” effects can occur without a conventional biochemical mechanism.

How drugs work: targeted enzyme inhibition

The professor explains that pain is linked to inflammation:

  • In the painful area, pain-inducing enzymes become activated.
  • These enzymes produce pain-causing substances that irritate nerves → pain.

Therefore, painkillers work by inhibiting the specific enzymes responsible for producing pain-causing substances.

Mechanism metaphor: Drugs are synthesized to match the 3D structure of target enzymes (like a “perfect fit”). Once bound, the drug blocks the enzyme’s function, reducing pain.

Side effects: on-target vs off-target binding

Even with “perfect” design, side effects arise because:

  • There are similar-shaped enzymes elsewhere in the body.
  • The drug may bind to unintended targets.

Technical framing:

  • On-target binding = intended correct match.
  • Off-target binding = unintended match due to similar structure.

Analogy: A Phillips screwdriver might fit, but if it slips into a slightly different/incorrect shape, it can cause problems.

Where drugs come from: evolution of drug material/technology

Historically:

  • Natural product medicines → chemical drugs (examples: aspirin, Tylenol).

Modern direction:

  • Rapid biotech advances lead to biopharmaceuticals:
    • Proteins, genes, cell therapies, stem cells
    • Vaccines using mRNA, etc.

The video frames medicine across “generations” of materials/approaches (discussed as 1st/2nd/3rd generation), suggesting current/future trends are biomedicine.

Why “protein medicine” is central

Protein is portrayed as the basis of life:

  • Enzymes, hormones, muscles, organs, hair/nails—nearly everything biological involves proteins.

Central dogma / molecular biology flow:

  • DNA acts like a cookbook (blueprint).
  • Cells act like restaurants producing protein “food.”
  • The process gene information → RNA → protein is emphasized as the central principle of molecular biology.

Proteins as “weapons”: Drugs are likened to precision bullets/missiles that target specific problematic proteins/enzymes to improve efficacy and reduce collateral damage.

Protein abnormalities cause disease; drugs correct, block, or remove them

Examples include:

  • Cancer: enzymes meant to “stop” instead keep driving uncontrolled cell proliferation.
  • Mad cow disease and Parkinson’s disease: described as protein misfolding, producing toxic/accumulating forms and progressive damage (e.g., nerve cell death, memory loss).

Therapeutic strategy depends on the protein problem:

  • Normalize protein function (restore normal behavior).
  • Disable, break down, or protect the abnormal protein.
  • Different companies may pursue different approaches even toward similar goals.

How scientists determine protein structure

Protein structures are too small to see directly; key methods include:

  • X-ray crystallography
    • Crystallize proteins, use X-rays, infer structure from diffraction patterns.
  • Nuclear magnetic resonance (NMR)
    • Infer structure by analyzing how the protein behaves under controlled conditions.
  • Cryo/temperature-based imaging
    • Lower temperature to slow movement, capture multiple images, then reconstruct structure computationally.

Cost challenge: Structure determination is extremely expensive—often compared to the effort level of obtaining a Ph.D. for a single protein.

Breakthrough: AlphaFold

  • Uses AI and large datasets to predict protein shape from amino acid sequences.
  • The video notes AlphaFold’s prominence and its association with the 2024 Nobel Prize in Chemistry.
  • It emphasizes that AI predictions aren’t perfect; experiments are still needed.

Precision medicine: designing drugs that fit disease-specific protein shapes

With knowledge of mutated/abnormal protein structures, drugs can be designed to bind precisely and avoid normal proteins.

Example: a leukemia treatment drug

  • Targets a specific abnormal fusion protein created by combined genes.
  • The abnormal protein acts like an engine that never turns off.
  • The drug is designed to fit it precisely with minimal effect on normal proteins.
  • The result is described as turning a previously severe disease into a chronic/manageable condition with fewer dramatic side effects.

Historical first protein drug example: insulin

Insulin is presented as the first widely known protein drug:

  • Initially discovered by extracting insulin from cow/pig pancreas (early diabetes treatment; referenced with Dr. Ben and Dr. Ting, dating to 1923).
  • In the 1980s, genetic engineering enabled mass production by inserting insulin-related genes into yeast or E. coli for recombinant production.

Why injection is used: The video explains that oral protein drugs are broken down in the stomach into amino acids, losing functional protein form.

Drug cost and limitations

Even when development and efficacy succeed, biopharmaceuticals are expensive, partly due to manufacturing difficulty.

  • A cited example: a Pfizer-approved blood-flow treatment priced extremely high per injection (figures given as 4.2 billion won).

Pricing dynamics:

  • High manufacturing unit cost and limited patient base can reduce price drops.
  • Market size and competition influence affordability.

Biohacking concerns

Because some people cannot afford treatments, the video mentions biohackers attempting DIY biopharmaceutical production using online information.

The professor stresses major risks:

  • Safety is not guaranteed.
  • It involves genetic manipulation, which is regulated and tightly controlled.

Drug development pipeline and timelines

New drug development is described as a long process with three stages:

  1. Research: identify disease causes and targets.
  2. Development: search for drug candidates that normalize the target (progress from target → candidate).
  3. Regulatory approval:
    • Non-clinical testing (animal models) for toxicity/efficacy
    • Then clinical trials phases 1, 2, 3

What counts as a “new drug”:

  • Innovative drug: a completely new target; no prior equivalent exists.
  • Improved new drug:
    • First-in-Class: improved mechanism class (higher tier)
    • Best-in-Class: optimized improvements relative to existing drugs (lower tier)

Timelines:

  • Innovative/overall development typically takes 20–30 years from basic discovery to clinical impact (as described).
  • Even improved versions can require extensive trials; changing even a single part of a chemical structure may force rerunning certain trial phases.

Why scientists persist

The professor characterizes drug development as high-risk/high-return:

  • The “return” is saving patients who otherwise have no cure.

He also describes a long-term dream:

  • A future where diseases are predicted and prevented, allowing long healthy life with less need for drugs.

He hopes for novel delivery methods (example suggested): medicines absorbed through skin (e.g., “like a sticker”).

Methodology / instruction-like content (detailed bullet format)

Targeted drug design approach (as explained)

  • Identify the biological cause of symptoms (e.g., inflammation activating pain enzymes).
  • Determine the target enzyme/protein activated in the diseased/painful region.
  • Design/synthesize a drug that:
    • Has a 3D structure match for the target protein/enzyme binding site.
    • Binds so it blocks the enzyme’s function, preventing production of pain-causing substances.
  • After binding, expect symptoms to decrease/disappear because downstream molecule production is interrupted.
  • Reduce side effects by minimizing off-target binding:
    • Check for enzymes/proteins with similar shapes that could also bind the drug.
    • Aim for maximal on-target selectivity.

Protein structure determination workflow (conceptual methods)

  • Concentrate protein to high levels to obtain structural information:
    • Crystallize → use X-rays → infer structure from diffraction patterns.
    • Use NMR to infer structure from how the protein behaves.
    • Lower temperature to slow motion → capture multiple images/videos → reconstruct structure computationally.
  • If using AI prediction (AlphaFold):
    • Predict 3D structure from the amino-acid sequence.
    • Validate/complete via experiments, because AI may miss certain parts.

Drug development pipeline (three-stage process)

  • Stage 1: Research (university-led work described)
    • Analyze disease causes.
    • Identify targets.
  • Stage 2: Development
    • Screen/search drug candidates that normalize target function.
    • Once a candidate is chosen, it becomes the substance to test.
  • Stage 3: Preclinical + Clinical + Regulatory approval
    • Non-clinical trials in animal disease models (toxicity/efficacy checks).
    • Clinical trials:
      • Phase 1
      • Phase 2
      • Phase 3
    • Apply for regulatory approval before release.

Protein medicine feasibility rule (oral vs injection)

  • If the drug is a protein intended to function as a protein:
    • Oral administration usually fails because digestion breaks it down into amino acids.
  • Therefore:
    • Use delivery methods that preserve functionality (the video specifies injection for insulin-type protein drugs).

Speakers / sources featured (as named in the subtitles)

  • Kim Seong-hun — professor at Yonsei University College of Pharmacy (main featured speaker)
  • Dr. Ben — (named as “Dr. Ben” in the insulin history; associated with Toronto in the subtitles)
  • Dr. Ting — (named as “Dr. Ting” in the insulin history; associated with Toronto in the subtitles)
  • AlphaFold / DeepMind / Google DeepMind — referenced as the AI system and its research origin
  • Nobel Prize in Chemistry (2024) — referenced as an award associated with AlphaFold
  • Pfizer — referenced in the context of a high-cost approved treatment
  • United Nations / “Hoam Prize” (Hoam Medical Prize) — referenced as an award context (not a person)

Original video