This is going to be hugely dependent upon dosage. I assume by "active delivery" you mean targeting moeity and/or triggered endocytosis.

Your highest uptake levels will be from triggered macropinocytosis, at which point the cell will engulf lots of extracellular fluid. From there, the dosage is only limited by what concentration of particles is in solution, and the limiting of toxic effects due to membrane destabilization (depending on surface functionalization). 

You'll also have later toxicity, but in terms of sheer loading, membrane stability is your main hurdle. 

Adding to Kameron's answer, the number of intake also depends upon the available receptors in each cell for the targeting moiety to attach. Sometimes, if they bind only to the cell membrane, then more particles can go in while if the nucleus is targeted, only few can get inside (due to the nuclear signal may be?).

I work with rods and I have never come across works that have actually reported the number of particles entering the cell. 

Theoretically, using the concentration of AuNPs and the number of cells seeded, we can get a ratio of NPs:Cells. But when it comes to actual use, the number of NPs inside each cell is different.

Hi Serge,

The uptake of GNPs in cell follows of lognormal distribution with the cell volume.

You can have differences between the GNP uptake  by a factor > 10.

I have in my mind a maximum of 10 000 GNPs /cell.

The article : 

DOI: 10.1039/C3AN01406A

will be interesting for you 

Have a nice week end

Eric

I thank all for valuable answers! Indeed, the number of NPs will depend on the number of targeting sites (e.g., surface receptors that are (over) expressed on a surface of an individual malignant cell). Typical numbers for such receptors cover a broad range from 10^3 to 10^4 sites per cell. However, some receptors like PSMA are presented at very high numbers, almost 10^6 per cell making them an attractive target for molecular imaging of cancer. It's also understood, than an elementary act of functionalized Au NP uptake by an individual cell proceeds differently in vivo and in vitro because of the RES.

So, how good are we in targeting efficiency? Can we achieve 10^3, 10^4 or even 10^5 Au NPs per cell and still preserving the functionality of the live cell? Let's think about Au NPs as drug/contrast agent etc. that has to be delivered at as high as possible numbers per cell for diagnostic or therapeutic purposes that will not harm the cell. For the sake of discussion, I'm not considering bare Au NPs but those with polymer, small molecules, peptide, protein etc. coatings.       

As I have observed, the more number of particles inside the cell, the more stress for the cells and they start dying. This may be due to the unusual number of vesicles required to take up the NPs causing energy deprivation inside the cell (leading to apoptosis?). So, in my opinion less than 10^3 Au NP inside cell can provide good diagnostic and therapeutic use.