When planning experiments, you need to decide what “experimental system” you’re going to use – variants of in vitro, in cell culture, and in vivo systems. These span from: the most minimalist, deconstructionist approach that tries to build things from scratch to get at core mechanisms while allowing for the highest control – to experiments done in whole organisms that aim for greatest realism – as well as experiments in cell culture that fall in between the two extremes. There are different reasons to choose different systems and different pros & cons of each. So here’s an overview of some things to keep in mind 

super overview of some of the systems I’m talking about and then details:

from most control to least control (and least realism to most realism)

in vitro (outside of cells) > in cell culture > in vivo (inside of organisms)

we can broaden that a little to encompass more systems…

reconstituted > lysate-based > in cell culture > in organoids > in tissue > in vivo

in vitro – outside of cells, “in a test tube*” – most experimental control but least realistic – best for dissecting mechanisms – how do things work?

  • *doesn’t have to literally be in a tube, just outside of a cellular and organismal context
  • most experimental control
    • can easily introduce labeled components 
    • can easily introduce compounds like initial hits for pharmaceutical drugs (pre-optimization) that might not be easily cell-permeable (can’t get into cells if you just bathe the cells in them) and/or bioavailable (won’t get into cells if you give the drug to animals)
  • decisions to make…
  • reconstituted vs. lysate based
    • reconstituted: you make it “from scratch” purifying the individual components (if necessary) and mixing them together, typically using only what is necessary and sufficient 
      • pros: most control
        • can easily alter concentrations of different components individually
        • can easily introduce mutated variants without having to do genetic engineering in cells or anything
        • best consistency – don’t have to worry about batch-to-batch, cell line-to-cell line, passage to passage, animal to animal… variability
      • cons: 
        • least realistic
          • you might be missing components that aren’t *necessary* for minimal function but help a reaction happen in vivo
        • lots of components
          • might have to purify a lot of different components
            • recombinant expression (e.g. stick gene in bacteria and get bacteria to make it for you) and in vitro transcription (making RNA in a test tube) has made this an easier task than if you had to purify native proteins, but there might be a lot of components you need to express and purify
    • lysate-based: you break cells open (lyse them), remove the membrane bits, and use the remaining lysate (un-purified or partially purified) as a reaction component – along with additives particular to your experimental question 
      • pros:
        • more realistic than with reconstituted
        • might contain components that you aren’t even sure what are but they enable and/or enhance the reaction you’re trying to study
        • you can still add components that aren’t present or add more of components that are present but not in high enough concentrations for your purposes
      • cons:
        • less control than with reconstituted
          • don’t even know precise concentrations of things
          • you can’t change components that are already present in the lysate unless you change the cell type or somehow remove the components that are in the lysate (e.g. by immunoprecipitating out the wild-type (“normal”) version of a protein to test its importance and/or replacing it with a mutant form you expressed & purified separately)
        • batch-to-batch variability
  • least maintenance required – getting it up and running can be daunting, and there’s a large up-front investment, but if you’re able to purify enough of the components to last a long time (and they store stably enough) you can do a lot of experiments without having to do any more work (no having to feed and passage cells, etc.)
  • quickest ability to test questions (once the system is working)
  • least realistic
    • may be missing components
    • don’t have compartmentalization that may be important inside of cells (in cells, for instance, scaffolding proteins and/or functional RNAs often play a role in keeping components of a pathway or reaction nearby one another (and thus increasing their local concentrations))
    • if you want to study membranous proteins/systems, you may need to consider things like liposomes (artificial lipid bubble things)
    • results might not be physiologically-relevant
      • e.g. if you’re testing a drug that can’t even make it into cells, and/or that get quickly metabolized (altered and/or broken down) in the body what should you make of the results? 
  • decisions to make:
    • reconstituted
      • what are you going to make yourself vs. purchase?
        • what is commercially available?
          • is it simple enough to make yourself?
            • is it cheaper to make yourself? taking into account amount you need, time & energy of making it
      • what organism’s components are you going to use (e.g. E. coli aminoacyl tRNA-synthetases (AARSs) or the “homologous” human versions?
        • note: often published protocols mix and match where components come from without really offering an explanation – sometimes it might be important (e.g. one species’ version is more efficient or works better in vitro – maybe it doesn’t require co-factors other homologs would), other times it might just be because that’s what they had on hand and/or that’s what people had used in the past and why fix what isn’t broken? But you might want to experiment with different sources and might have better luck sticking to a single organism where everything is maximally compatible with one another 
      • lysate-based
        • what type of cells are you getting lysate from?*
        • are there special things you need to do when harvesting them and preparing lysate to make sure lysate is “competent” for your reaction?*
        • * look to published protocols & follow them (at least to start with – make sure you can recreate what’s been reported before you try switching things up so you don’t have to do more troubleshooting than necessary)

for a real-life example, and one that I’m currently thinking about lots, see the post on cell-free protein expression, where options include reconstituted approaches (e.g. the PURE system which uses purified E. coli translation components) and lysate-based approaches that use a cellular lysate (e.g. rabbit reticulocyte lysate, wheat germ extract, HeLa cell lysate) supplemented with amino acids and an energy-regeneration system. 

in cell culture* – medium level of control & realism, more variability

*sometime people refer to this as “in vivo” but, if you’re working with single cells, it’s only really in vivo if you’re working with a unicellular organism like bacteria or yeast

  • medium level of control
    • can introduce compounds more easily than in vivo, but less easily than in vitro
      • don’t have to worry about it getting into & out of the bloodstream, metabolism by the liver, etc. but may need to worry about cell permeability – can it get through the cell membrane?
      • can transfect nucleic acids (plasmids, mRNAs, siRNAs, etc.) using techniques you can’t use in vivo (at least not easily) – directly treating with transfection reagents, electroporation, gene genes, etc. 
    • can’t tightly control concentrations of reaction components
    • harder to change what’s already there
      • can use gene editing techniques to generate cell lines with “permanently” missing and/or altered versions of proteins, functional RNAs, etc. 
        • time-consuming & requires stringent validation, making sure you don’t have off-target effects with CRISPR messing up genes you didn’t want it to, etc. 
      • can use gene silencing techniques like siRNA-mediated RNAi to temporarily stop the expression of proteins of interest
      • can transfect instructions for making different proteins
        • can use some viruses or CRISPR to make it so those cells pass the instructions on to their progeny for longer-term instructions
  • medium level of realism
    • more physiological (biologically realistic) concentrations*
    • all* intracellular components necessary
    • *expression of various proteins & concentrations of various components may differ from cell type to cell type – and may further differ depending on cellular stresses, nutrient availability, etc.
    • lack signaling coming from other cells, extracellular matrix (ECM) etc. 
  • decisions:
    • what cell type(s)?
      • different immortalized cell lines are often used – keep in mind that these cell lines often come from cancer cells and may contain many mutations, chromosomal abnormalities, etc. -> you often need to validate findings in multiple cell types to make sure your results are more widely applicable 
      • sometimes patient or experimental animal’s cells can be grown in cell culture &  tested to study impacts of mutations
      • iPSCs (induced pluripotent stem cells) -> these are cells taken from organisms (e.g. a patient’s skin cells) and reprogrammed back into a stem-cell-like state where you can then add different mixtures of hormones & chemicals to coax the cells into taking alternative fates (differentiating into a different cell type) -> for example you can take a skin cell, make an iPSC from it and then differentiate that into a neuronal-like cell (type of brain cell) that contains the genetic makeup of the patient (note: I know from a little experience in undergrad that this is definitely easier said than done)
      • note: some cell types require more maintenance than others
    • what expression scale? wells, plates, flasks?

in vivo – inside of organisms – least experimental control but more realistic*

  • least experimental control
    • natural variability between organisms that may influence results, for better or for worse
    • have to worry about test compounds surviving metabolism & getting into the cells that need them and to the things in the cells you want them to react with 
    • changing things genetics-wise is even harder than in cell culture
  • most realistic*
    • * at least for the organism studied – but are the findings directly transferable to the system you care about? (i.e. a mouse isn’t a human)
  • depending on the organism, probably most expensive, likely longest time course, highest maintenance, may require institutional review board (IRB) approval – if you’re testing animals, you have to work with animals
    • note: if you’re working with bacteria or yeast these things don’t apply, and with plants and some model organisms like worms & flies things are much simpler than if you’re working with mice or monkeys, etc. I personally leave that work to others and am most comfortable in vitro.
  • can be hard to get a large enough sample size to make firm, statistically-sound, conclusions 
  • may be limited in the kind of data you can collect – is it possible to take samples over time? Or monitor a reporter (e.g. GFP expression) over the organism’s lifespan?

There are also systems that bridge the divide between cell culture and in vivo. For example, experiments can sometimes be performed in tissues and organs “ex vivo” (that have been taken out of the body of an animal for better control and/or easier measurements). Nowadays, cells can be used to grow “organoids” which are synthetic masses of cells that can have different cell types, etc. and try to recreate a more realistic cellular environment, including extracellular matrix (ECM), etc. These organoids are often made from patient cells. 

No matter what your system is, you’ll need to think about how you’re going to read out results. This will depend on your system and the questions you’re asking. 

Leave a Reply

Your email address will not be published.