Line Up: recommendations for the best human cell line

Here is a summary of the info I gleaned from asking for recommendations for the best human cell line. These were my criteria:

For context, we currently use a number of human cell lines in the lab: HeLa, RPE1, HCT116, SKOV3; as well as many others in the past: HEK293, DLD-1, U2OS.

I consider HeLa to be the almost perfect cell line. They are nice for imaging (great for biochemistry too), easy to culture and transfect… unfortunately they have a weird karyotype, so gene editing is painful. Plus, there’s always pushback from people who say HeLa are strange and not very representative of a “real cell”… a discussion for another day perhaps. So what cell line could be better?


Retinal Pigment Epithelial cell line (hTERT immortalized)

Nishit Srivastava @Nishit24

maybe RPE1

Dr Rafal Czapiewski @RafalEstrella

RPE1 are my favourite now!

Fernando M. Valbuena @FernandoMValbu3

I use RPE-1 cells regularly. They’re fantastic for imaging, aren’t tumor derived, and have a stable karyotype. I use Neon electroporation for transfection with very high efficiency even when using mini prepped DNA. Electroporation is a better method than lipid-based reagents.

Ewan MacDonald @EwanMac10292433

Have you tried RPE1 cells?

Andrii Gorelik @AndriiGorelik

With PolyFect you can get >50% transfection efficiency in RPE1

Miguel Rz @MickeRzO


Dr CJ Houldcroft @DrCJ_Houldcroft

What about TERT’d fibroblast lines?

Dr. Ceri Fielding @CeriFielding

Not that easy to transfect by conventional means, unless you’re transfecting siRNA.

Stephen Terry @Stephen17822072

Hi Steve I think rpe-1 htert cells I got are human near diploid flat so great for imaging transfect well for dna /siRNA the ones I have are puro resistant only drawback

Francesca Bottanelli @Fran_Bottanelli

We successfully made KIs in MCF10A and RPE1 cells. They transfect well with electroporation (NEPA gene).


Retinal Pigment Epithelial cell line (spontaneous immortalized)

Costin Antonescu @CostinAntonescu

We use ARPE19 for many applications (different from RPE1). Great for membrane traffic imaging especially TIRF. Reasonable for transfection, but drawback is they don’t edit easily.

Daniel Chauss @dchauss

ARPE19 also has reported karyotype heterogeneity


Human foreskin fibroblast

Tim Feinstein @TimNFeinstein

HFF cells have a lot of good properties and are more ‘stem-like’ than most cultured lines, but if you don’t like the transfection properties of RPE-1 cells then they might not work.

Caroline Clucas @Caroline_Clucas



Fetal lung fibroblast

Ishier Raote @ishier_raote

WI-26 or WI38?

Julian Nüchel (PhD) @JNuchel

That’s what I would suggest! My favorite cell line! Perfect for endocytosis, secretion and other intracellular trafficking pathways! Generated plenty of KOs and OE cells. Transfection worked best for me with Xtremegene HP from Roche/Merck

U2OS & COS-7

Bone sarcoma

African green monkey kidney

Hoang Anh Le (aka Anh), PhD @AnhHLe2702

U2OS or COS7 (monkey but human siRNAs and CRISPRs work for these cells)

Andy Moore @aaandmoore

I’ve had mixed success with human siRNA/CRISPR in COS7.  I guess it’s dependent on the specific gene.  I’ve switched to just designing things specifically for vervet, but then it’s kind of annoying to have a tool that only works for Vero/CV1/COS.

Hoang Anh Le (aka Anh), PhD @AnhHLe2702

I had one instance that the target siRNA didnt work for one gene in COS7, but luckily for other genes they worked. So yes, it is gene dependent.

Hoang Anh Le (aka Anh), PhD @AnhHLe2702

I see. I did have a bit of problem editing U2OS, but easier for COS7. I also heard SKOV3 cells are really easy to culture too. My labmate used SKOV3 and they seem to love them.

Jonny Nixon-Abell @AbellJonny

They aren’t I’m afraid…which is a shame because damn they look good! U2OS often aren’t diploid either but more stable karyotype. Could use the KOLF2.1 iPSC line, dont image so well, but meet your other criteria.

John Petrini @JohnPetrini

I would strongly advise against U2OS.  They are extremely unstable karyotypically.  See:  PMC3400581

Tyson V. Sharp @tyson_sharp

U2OS? Massive nuclei and massive cytoplasm!

Jullien Denis @JullienDenis1

Try U2OS

Stem cells

David Drubin @DavidGDrubin

For the reasons outlined here ( consider  WTC human stems.  Great protocols and a large catalog of engineered cell lines available thanks to @AllenCellInst. You can differentiate these cells into many other cell types (e.g., fibroblasts, cardiomyocytes)

Dr Agathe Chaigne @AgChaigne

Human, not mammals? mESC are pretty stable karyotypically (am I making up a word?) and I see a lot less division defects than hESC. They are not that hard to culture/transfect/image.

Max Gutierrez @maxgabgut

Cell line and karyotype stable are not compatible I am afraid. As suggested by others iPSC are an option that will not solve all your problems. If you use them at the iPSC stage results can be different from a fully differentiated cell. Good luck!



Graham Dellaire @DocDellaire

HT1080 are great, big flat cells for microscopy. They are not diploid but have a stable phenotype from the point of view of rearrangements but do spontaneously become tetraploid if abused.

Francesco Pasqualini @fspasqualini

We love them but they also do like to crowl on top of each other which is not ideal for all microscopy needs

Aoqun James Song @AJamesSong

How about HT1080?


Triple-negative breast cancer TNBC

Aurélie Dobric @AurelieDobric

I may be too late for answer but I haven’t seen one cell line in the answer below : MDA-MB231. They are gorgeous for live imaging.

Shem @Shem_Darius

Not sure it checks all those boxes, but look at mda-mb-231 cells if you’re into endosomes. They’re huge!

Sherif Ibrahim. MBBCH, MSc, PhD @444sherif

Some authors reported difficulty in their transfection, but they worked well in my hand. I am not sure about their karyotype but they are easy to culture, and I believe they are great for microscopy.


Colorectal cell line

Janardan Gavade @janardangavade

Maybe Dld-1 cell line can be the answer!

Melinda Beccari @MelsNeurons

Many here in lab use DLD-1s! If I’m not mistaken they fit all the items on your list. Let me know if you guys have questions about them or want a vial. We’ve used SH-SY5Y and U2Os also but always with lentivirus or nucleofection, not transfection.


Breast cancer epithelial

Valentine Comaills @ValComaills

MCF10A, great to work with. Media a bit weird but genomically stable.

Francesca Bottanelli @Fran_Bottanelli

We successfully made KIs in MCF10A and RPE1 cells. They transfect well with electroporation (NEPA gene).

THERY Manuel @ManuelTHERY

and hard to transfect

Valentine Comaills @ValComaills

Lentivirus, retrovirus and siRNA were fine for me.. but nothing too agressive as they are ‘normal’ cells and don’t like to be messed up, and end up to sense it and stop to grow.

HCT 116

Colorectal cancer

대학원생 달토끼 @BunnyTheHuman_

How about HCT116?

ChrisStaples @StaplesLab

As well as the mutator phenotype, they are not so beautiful in fluorescence microscopy.

대학원생 달토끼 @BunnyTheHuman_

Although its small size, we use hct116 for PALM imaging.

Graham Dellaire @DocDellaire

I would stay a million miles away from HCT-116 due to MLH1 mutations and hyper mutation phenotype. They are a different cell line every few passages. Nearly every gene can accumulate point mutations.

Graham Dellaire @DocDellaire

For future reference (see Davis et al., 1998 Cancer Res.) since both Cellosarus and Wikipedia fail to inform users that HCT116 are MLH1-deficient…


Lung fibroblast

Fausto Capelluto @FaustoCapelluto

Have you tried MRC5? In my experience, they are very good for imaging and easy to culture.

Samantha Lewis @SamLewisPhD

MRC5 have beautiful mitochondria but ER looks like cotton candy… IME

Only the lonely: lines with just one recommendation


Lung cancer

SA @SSA07844173

How about A549 cells?


TNBC cell line

Jez Carlton @JezCarlton

I use CAL-51, near diploid, transfect well, edit well, great for imaging, divide every day. TNBC line with I think activating PI3K mutation.


Intestinal epithelial

Dave Boucher @dave_boucher

Hiec-6 cells . Intestinal epithelial cells, grow well , easy media (optimum 4%fcs + egf)



Polina Lishko @PolinaLishko

HaCaT (human keratinocytes) are great for 1-3, but they are aneuploid. Not sure how stable their karyotype is.


Lung fibroblast

Samantha Lewis @SamLewisPhD

IMR90. Normal karyotype, easy to transiently transfect, and high editing efficiency. They do senesce after 15-20 passages though.


TNBC cell line

Jeeyun Chung @JeeyunC

We use SUM159 breast cancer cell line. Diploid, easy to KI and KO, okay to transfect, beautiful ER structure for imaging!

Ewan MacDonald @EwanMac10292433

We use SUM159. They have nice trafficking, near diploid.



Tim Triche, Jr. @timtriche

Presumably HEK293T is no good for microscopy of the sort you’re doing? Otherwise it’s nice



Candice @candice_ab

Jurkat cells?


Inevitably the answers were a classic case of YMMV. One person’s idea of a cell line that’s nice for imaging can be quite different from someone else’s. On top of all this, a wonderful cell line which doesn’t express the protein you are interested in (or do the thing you work on) is no use to you at all.

Does the perfect cell line exist? The strongest contenders all had drawbacks.

  • RPE1 were the most popular suggestion. They’re great but hard to transfect.
  • U2OS great but karyotypically unstable.
  • HCT116 again the genomic concerns raised by Graham are off-putting.

There were several lines that we haven’t used and I’m keen for us to try. CAL-51, HT1080 and stem cells are all on our list.

Thanks to everyone who replied, I learned a lot.

The post title comes from “Line Up” by Elastica which opens on their eponymous debut LP.