NIAID IN COLLABORATION WITH COLORADO STATE UNIVERSITYViruses infect just about every living organism, be it man, mouse, flea, or bacterium. These parasites cannot reproduce in isolation: they need to get inside the hosts’ cells. That’s why virologists need cell cultures, but to wield those cultures well they must understand both viruses and host cells.
It’s not as simple as tossing the two together in a flask or petri dish, notes Charu Kaushic, a professor at McMaster University in Hamilton, Ontario. As a postdoc, she studied the innate immune system using epithelial cells from the human female reproductive tract. When she started her own lab, Kaushic decided to investigate how the sexually transmitted viruses HIV and herpes simplex 2 interact with those same cell types. Establishing the cell culture system—completely characterizing the cells, working out viral dosing and readouts, and achieving reproducible,...
There are several reasons virologists culture cells, says Marshall Bloom, associate director for science management at Rocky Mountain Laboratories, a division of NIH’s National Institute of Allergy and Infectious Diseases located in Hamilton, Montana. Clinical virologists might add a patient sample to cells, looking for evidence of infection. Researchers also use cells as biological test tubes to grow viral stocks. Moreover, they infect cells with viruses, or express individual viral proteins, to follow the virus’s actions and the host cell response. “Cell cultures have played a critical role in modern infectious disease research, particularly in the area of viruses and the expression of viral gene products,” Bloom says.
Here, The Scientist examines the decisions virologists must make, and techniques they can apply, as they design virus–cell culture systems.
The cells
The type of cells chosen depends on the virus in question. Finding the right host can be a bit of a “detective game,” Bloom says. The obvious place to start is with cells that match the animal and tissue that naturally host the virus. Camel cells would be a first choice for Middle East Respiratory Syndrome coronavirus (MERS-CoV), for example.
But that strategy doesn’t always pan out. When Bloom was studying Aleutian mink disease virus, he tried mink cells, plus every other mammal line he could purchase. The only ones he managed to infect were cat kidney cells (J Virol, 73:3835-42, 1999). The researchers always had to keep in mind that they were studying the virus in an atypical host.
If a virus turns out to be picky, several cell lines, such as HeLa, welcome a variety of viruses, Bloom says. Baby hamster kidney (BHK) cells and fibroblasts from chick embryos are also a good bet, says Richard Condit, an emeritus professor at the University of Florida in Gainesville.
In addition to a cell type’s origin, virologists must choose between primary cells cultured directly from an organism, or immortalized cell lines that can or have been passaged for years. Each option has advantages and disadvantages.
Immortal cell lines such as HeLa are convenient, easy to grow, and highly reproducible, due to their clonal nature. The majority of virus studies use such lines, says Mohsan Saeed, a postdoctoral associate at Rockefeller University in New York City and at the Center for the Study of Hepatitis C, a collaboration of Rockefeller, Weill Cornell Medical College, and New York–Presbyterian Hospital.
However, such cells bear only a passing resemblance to their counterparts in a whole organism. They divide constantly, altering their metabolism. They often dedifferentiate, regressing to a primitive state. Crucially for virologists, cell lines tend to mount an abnormal immune response, so they may not defend themselves against viruses as cells would in vivo. Results from cell lines can be inconclusive or just plain wrong, says Vyas Ramanan, a graduate student in the Laboratory for Multiscale Regenerative Technologies at MIT.
For these reasons, Saeed recommends confirming results in primary cells when possible. Primary cells look and act more like cells in vivo, but have their own complications. They require special skill to cultivate, and will eventually die out.
If a lab wants human primary cells, acquiring them may add another layer of complexity. Often the desired cell types are available from commercial vendors: for example, 21 families of human primary cells are sold by Lonza, including hepatocytes ($500 for an ampule containing 3–6 million cells). However, sometimes researchers need a tissue type that no company offers, or want to select exactly who the donor is, so they have to go straight to the source by contracting with hospitals.
For example, Kaushic’s group asks women who have healthy uteruses, but undergo hysterectomies for reasons such as excessive bleeding during menopause, to donate their tissue. The scientists try to process the samples as soon as the hospital’s pathologist has deemed them normal. “We are on call all the time,” Kaushic says. (See “The Spleen Collectors,” The Scientist, August 2015.)
Unlike the identical cells in cloned lines, donor tissues vary as much as people do. That can be a plus or a minus. Kaushic prefers the assortment because she wants to understand how viruses infect all women. However, those doing molecular virology might find that using primary cells from different batches makes their results less reproducible.
The virus
COURTESY OF MOHSAN SAEEDThe cells, of course, are only one half of the culture system; viruses have their own version of the line-versus-primary question. Once viruses start growing in culture, they adapt and may acquire mutations that alter growth rate or virulence. In Bloom’s case, the Aleutian mink virus grown in the cat kidney cultures caused fairly mild disease in minks, and he was never able to figure out why.
The choice of a virus source depends on your goals. If you want to infect every single cell in a culture and analyze the cell response, a highly infectious lab strain might be a good way to go, Saeed says. However, to study the effects of a potential treatment, he prefers a wild-type virus.
After selecting cell and virus types, scientists must address basic questions about how much virus to use, and how long to let the infection run. “These are not the world’s most exciting experiments, but these are the meat-and-potatoes experiments with which you make a good recipe,” Kaushic advises. The classic technique to quantify a viral stock is to infect a culture, and count how many empty spots, or plaques, appear on a monolayer of cells grown in a petri dish or flask. Each plaque corresponds to infection by a single viral particle.
With that number in hand, researchers can achieve their desired multiplicity of infection (MOI), the virus:cell ratio in an infection experiment. An MOI of 1, with one virus for every cell in the dish, will only infect about two-thirds of the cells, says Saeed. Because the virus lands randomly, some cells will get none, while others will be invaded by two or three particles. To infect every cell in a culture, he recommends an MOI of 5. To study how a virus transmits between cells, he suggests an MOI of 0.01, so only one in 100 cells is affected at first, and then those cells release virus to infect their neighbors.
Researchers also need to monitor infection. The most obvious change is called the cytopathic effect, or CPE. While most cultured cells grow in flat monolayers, infected cells may round up, fuse together, or burst. “The cells can look pretty awful, and still be alive and producing virus,” says Condit, who authored a chapter on virus cultivation in the manual Fields Virology.
Not all infections lead to CPE, however, so you may need to employ other techniques. For a population in a dish or well, researchers can perform PCR to identify viral genes, or use Western blotting to detect viral proteins. Kaushic cautions that finding viral nucleic acid or proteins does not necessarily confirm the presence of viruses capable of infecting and replicating. For example, many HIV researchers measure the presence of the core viral protein p24. If they are not careful, she says, they could be identifying p24 from the initial inoculum they added to the cultures, or from incomplete, inactive viral particles.
Kaushic prefers to assess viral replication in multiple ways. For example, researchers in her lab add media from their experiments to an HIV indicator cell line. These cells, derived from the HeLa cell line, express receptors for HIV as well as a gene for β-galactosidase controlled by HIV sequences. When exposed to live HIV, they turn blue in the presence of X-Gal, a dye-modified analog of lactose. The “TZM-bl” indicator cells are available for free to noncommercial researchers from the NIH’s AIDS Reagent Program and commercially from ATCC for $200.
Other techniques allow researchers to identify individual cells that are infected. These include fluorescence in situ hybridization for viral genes in live cells, or immunofluorescence for proteins in fixed cells (see “‘Alive’ and in Focus,” The Scientist, October 2012).
COURTESY OF CHARU KAUSHIC AND SARA DIZZELLSome labs insert the gene for a fluorescent protein into the viral genome, so infected cells in living cultures will glow. However, the extra gene may slow down the viral life cycle, Saeed says. The additional nucleic acid can alter the secondary structure of RNA, slowing down translation, and the cell might be inefficient at stuffing the longer genome into virus particles.
Alternatively, researchers can alter cells so they indicate when they’ve been breached. The Rockefeller lab did this for a human hepatoma cell line, HuH-7.5, that they infected with hepatitis C. The viral protease cleaves a cellular protein called MAVS, normally found anchored in the mitochondrial outer membrane. The researchers created a chimeric gene, linking the cleavage site and mitochondrial-targeting sequences of MAVS to the code for red or green fluorescent protein as well as to a nuclear localization sequence. In uninfected cells, the resulting protein remains tethered to mitochondria. When the virus enters a cell, the protease cleaves the reporter protein, and the nuclear localization sequence takes over. The researchers can easily differentiate uninfected cells with unlit nuclei from infected cells with glowing nuclei. (Nat Biotechnol, 28:167-71, 2010)
Tissue engineering
Sometimes, the simple recipe of one cell type plus virus is not enough to model infection. “For certain viruses, the ecosystem that you culture them in is very important,” Ramanan says. For example, hepatitis viruses do not interact only with hepatocytes in vivo. The liver contains endothelial cells, immune cells, and other cell types that also influence the process of viral infection, even though those cells are not infected. This has led Ramanan and colleagues to explore more-complex culture systems (reviewed in Annu Rev Virol, 1:475-99, 2014).
Researchers in the MIT lab developed a micropatterned hepatocyte culture system, using stencils to apply dots of collagen into the bottom of each well. They then seed primary human hepatocytes, which selectively adhere to the collagen islands and surround the seeded islands with a sea of supportive mouse embryonic fibroblasts. The liver cells maintain their polarization and survive for weeks in culture, compared to the few days they would last on their own, while supporting the life cycles of hepatitis B and C (Nat Biotechnol, 26:120-26, 2008; PNAS, 107:3141-45, 2010).
Another way to make cultures more like real tissues is to grow cells in three-dimensional substrates (see “Enter the Third Dimension,” The Scientist, September 2012). For example, researchers can combine different cell types in a gel-like matrix to create three-dimensional “organoids.” (See “Orchestrating Organoids,” The Scientist, October 2015.) However, for virologists, there is a drawback. Depending on the matrix used, the virus may or may not traverse it, and it’s hard to say how easily the particles can reach cells in the organoid’s interior. “You don’t have as much control as you have with a monolayer,” Saeed says.
Tissue engineering for virology is in its infancy, and for many questions cell monolayers will work just fine, Ramanan says. “You only want to add as much complexity as you need to answer the questions that you’re asking.”