Cell biology / mechanobiology
Cell mechanobiology: how cells sense gravity and load
Cells sense load through the actin cytoskeleton, focal adhesions and the nucleus. When that load changes, their gene expression and behaviour change too. On Earth this matters for bone loss, blood vessel disease and ageing. In space, bone and blood vessel cells lose their normal load. Flight studies show actin changes within seconds of microgravity, and changes in adhesion, stress and metabolism genes after days. Only free fall removes the gravity load for days. Ground devices such as clinostats only turn the gravity vector, and in one osteocyte study they did not reproduce the flight changes. An on-board 1 g centrifuge gives a control that shares launch, radiation and hardware, so the effect of gravity can be separated from the other spaceflight factors.
Who it is for
- University mechanobiology and cell biology labs (cytoskeleton, focal adhesion, nuclear mechanics, YAP/TAZ signalling)
- Space agency life-science programmes studying astronaut bone loss and cardiovascular deconditioning
- Pharma and biotech groups working on osteoporosis, bone anabolic drugs and vascular disease who want an unloading model
- Partial-gravity research for Moon and Mars missions (the same centrifuge can run 0.16 g or 0.38 g)
- Makers of tissue-engineering scaffolds and live-cell reporters who need flight validation
What earlier flights found
| Year | Mission | What flew | What it found |
|---|---|---|---|
| 2013 (published) | ESA SPHINX (Spaceflight of Human Umbilical Vein Endothelial Cells: an Integrated Experiment), Progress 40P missionISS | Human umbilical vein endothelial cells (HUVECs); 12 in-flight and 12 ground-based control modules; 10 days | Post-flight microarrays found 1023 significantly changed genes, mostly in cell adhesion, oxidative phosphorylation, stress responses, cell cycle and apoptosis. TXNIP was the most up-regulated (33-fold); HSP70 and HSP90 the most down-regulated (5.6-fold). Focal adhesion genes were widely affected. IL-1 alpha and IL-1 beta secretion increased; nitric oxide did not change. The control was on the ground, not an in-flight 1 g centrifuge.1 |
| 2015 | TEXUS 52 sounding rocket (with the 24th DLR parabolic flight campaign)Sounding rocket (about 6 minutes of microgravity) and parabolic flight | First flight of the FLUMIAS compact fluorescence microscope; FTC-133 human thyroid cancer cells expressing Lifeact-GFP to show F-actin live | The authors report it as the first in-flight live-cell imaging of human cells. On the rocket, the actin cytoskeleton changed rapidly after entering microgravity. Controls were 1 g and 1.8 g (short-arm human centrifuge) runs on the ground; there was no in-flight centrifuge.23 |
| 2018 | TEXUS-54, second pioneer mission of the FLUMIAS microscopeSounding rocket | FLUMIAS confocal spinning-disk fluorescence microscope; primary human macrophages with four live dyes (Nuclear Violet, Calcein, LysoBrite, SiR-actin) for nucleus, cytoplasm, lysosomes and actin | In 1e-4 to 1e-5 g for about six minutes, compared with 1 g controls, the team measured changes in 3D cell volume and surface, nuclear volume and the actin cytoskeleton. SiR-actin signal dropped significantly after 4-19 s of microgravity, then adapted until 126-151 s.45 |
| 2006 | LYCIS T-cell activation experiment in the KUBIK facility, Soyuz 13S (TMA-09)ISS | Human T cells activated for 1.5 h at 36.5 C; 1 g samples spun in the central positions of the KUBIK centrifuge during flight; stopped with RNAlater | With an in-flight 1 g control, microgravity inhibited Rel/NF-kB pathway activity and immediate early gene transcription in T cell activation; TNF was about 15-fold lower than in the 1 g samples. It is an example of how an on-board 1 g control can be run. T cells are immune cells, so it is cited for the control design, not as a mechanobiology result.6 |
| 2015 | SpaceX Dragon-6 (CRS-6) resupply mission to the ISSISS (cells flown up on a resupply mission) | Murine Ocy454 osteocytic cells on a porous polystyrene 3D scaffold, given 2, 4 or 6 days of microgravity and compared with 1 g cells on Earth | Microgravity impaired osteocyte differentiation and changed glycolysis pathways. A set of mechanically sensitive genes was consistently regulated in several cell types exposed to microgravity. Ground runs in the NASA rotating-wall-vessel did not adequately reproduce the flight changes.78 |
How LELP-1 runs it
Sample. Adherent human endothelial cells (primary HUVEC or an immortalised endothelial line) carrying two live reporters: Lifeact-GFP for F-actin and H2B-mCherry for the nucleus. Option: swap in an osteocyte line (for example Ocy454) for a bone customer. One frozen cell lot feeds the flight 0 g bay, the flight 1 g centrifuge bay and the ground bays. Container. 0 g bay: 24-well glass-bottom cassette, fibronectin-coated, with gas-permeable lids, in the sealed BSL-2 bay. 1 g bay: a 24-well ring cassette on the centrifuge rotor, all wells at one 8 cm radius with well floors facing outward. Each bay has its own -80 C passive cryo cassette with cells until thaw. Reservoirs: HEPES-buffered growth medium, wash buffer, 4% paraformaldehyde (PFA), PBS, RNAlater-type stabiliser, waste. Launch. Cryopreserved cells (-80 C passive cassette); media and fixatives at 4 C bay setpoint. Light and gas. No growth light needed. The bay stays dark except for short microscope LED exposures, kept low to limit phototoxicity. No CO2: 25 mM HEPES-buffered medium with gas-permeable lids and the sealed bay air headspace.
- day 0
Launch with cells frozen at -80 C and fluids at 4 C. In orbit: bay checkout, seal and pressure check, dosimeter on. Spin up the 1 g centrifuge bay and confirm 1.0 g at the ring radius from rotor speed and its accelerometer.
- day 1
Set both bays to 37 C. Thaw cells automatically, dilute out the cryoprotectant, seed into the fibronectin-coated wells. In the 1 g bay, cells are delivered through the rotary fluid coupling and attach on the spinning rotor. In the 0 g bay: brightfield + GFP + mCherry images every 1 h for 12 h, then every 3 h.
- day 2
Replace medium to remove cryoprotectant and unattached cells. Keep time-lapse imaging. Hold robotic arm moves outside imaging windows and log every arm move against well images.
- day 3
Early end point: fix 8 wells per bay with 4% PFA for 15 min, then flush to PBS. The 1 g wells are fixed while still spinning.
- day 5
Medium change. Continue time-lapse imaging of the live wells: cell area, shape, actin stress fibres, nuclear area and shape.
- day 7
Last live images. Move conditioned medium to a reservoir. Fix 8 wells per bay with PFA then PBS; stabilise 8 wells per bay with RNAlater-type reagent. Fix and stabilise the 1 g wells before stopping the rotor. Then set both bays to 4 C.
- day 8
Robotic arm takes the fixed 0 g bay to the central 40x fluorescence microscope for higher-resolution images of actin and nuclear shape, then returns it. Repeat for the stopped, fixed 1 g bay. Hold at 4 C until de-orbit.
- day 9
De-orbit, splashdown, bays removed. Samples handed to the customer courier about 20 min after splashdown and carried at 4 C to the lab (about 1-2 days).
The payload module, in CAD
Every item in the module sheet, drawn and packed into the large bay deck by deck. The back face mates with the bay: blind-mate connector, and a cold wall against the bay's thermoelectric block. The front face carries Dexter-L's grapple fixture, and the fluid lines run down a channel just behind it. What you load before launch is marked blue in the configurator's payload guide.
CAD generated by cad/build_modules.py (CadQuery, OpenCascade kernel) from the catalog's module items: concept geometry, not a released design. Standard parts keep their real formats; the rest are sized from their catalog mass. The front panel and grapple fixture are left off in the 3-D view and are in the STEP.
Checks
| Fits the large bay | FITS | 313 x 191 x 365 mm envelope; 58 mm of height to spare |
| Mass | AS CATALOG | 5.80 kg in the parts, 5.80 kg in the catalog |
| Parts' boxes | 40 % of the 21.1 L inside | |
| Housing | THIN | 0.80 kg of structure and containment: a 0.38 mm aluminium skin |
| Centre of mass | -7 mm across and -8 mm front to back (+ is toward the front) from the middle, 142 mm up | |
| Lines | 7 fluid lines and 3 harness runs, routed in the channel behind the front panel |
Inside, top to bottom
- Top deck controller and data storage
- Experiment deck imaging stack: inverted microscope, XY stage, sample; 2 micro-pumps, 4 valves on a manifold
- Floor deck fluid pack; -80 C passive cryo cassette
- The XY stage travels 96.5 x 57.9 mm, so every one of the 24 wells reaches the objective; with the stage the imaging stack is 232 x 151 mm, which sets the module's depth budget.
- Reagent volumes the catalog does not state are split evenly from 75 % of the reservoir item's mass. The waste bag holds the spent medium and washes; fixatives and stabilisers stay in the wells and come home.
- Fluid pack: 5 reagent bags in 2 layers (197 x 95 x 46 mm), every port on a front edge.
- The waste bag lies under the reagent bags and fills the room they free: they hold 525 mL, it needs 315 mL.
- The catalog gives 0.80 kg for structure and containment; as an aluminium box that is a 0.38 mm skin on 0.24 kg of posts and shelves. Thin: the bay rails have to carry the launch loads, or the module needs more structure mass.
- The grapple fixture stands 26 mm proud of the front panel, into the bay opening; everything else is inside the bay envelope.
Components (5) and how each catalog item was drawn
| Component | Deck | kg | Box mm |
|---|---|---|---|
| fluid pack | Floor deck | 0.70 | 197 x 103 x 51 |
| -80 C passive cryo cassette | Floor deck | 0.60 | 80 x 119 x 80 |
| imaging stack: inverted microscope, XY stage, sample | Experiment deck | 2.80 | 232 x 151 x 174 |
| 2 micro-pumps, 4 valves on a manifold | Experiment deck | 0.50 | 50 x 94 x 38 |
| controller and data storage | Top deck | 0.40 | 140 x 101 x 28 |
| Catalog item | kg | Drawn as |
|---|---|---|
| LumaScope-class inverted microscope (brightfield + 2 fluorescence channels) with XY stage and camera | 2.20 | microscope |
| 24-well fibronectin-coated glass-bottom cassette with gas-permeable lids | 0.60 | cassette |
| -80 C passive cryo cassette with cells | 0.60 | cryo |
| Filled fluid reservoirs (medium, wash, 4% PFA, PBS, RNA stabiliser, waste) | 0.70 | bags |
| Pumps, valves and manifold | 0.50 | pumps |
| Module controller and image storage | 0.40 | avionics |
| Structure, brackets and containment insert | 0.80 | frame |
Every part with mass (53)
| Part | Role | kg | Size mm |
|---|---|---|---|
| medium (105 mL) | fluid | 0.116 | 63 x 95 x 22 |
| medium port | fluid | 0.010 | 5 x 8 x 5 |
| wash (105 mL) | fluid | 0.116 | 63 x 95 x 22 |
| wash port | fluid | 0.010 | 5 x 8 x 5 |
| 4% PFA (105 mL) | fluid | 0.116 | 63 x 95 x 22 |
| 4% PFA port | fluid | 0.010 | 5 x 8 x 5 |
| PBS (105 mL) | fluid | 0.116 | 63 x 95 x 22 |
| PBS port | fluid | 0.010 | 5 x 8 x 5 |
| RNA stabiliser (105 mL) | fluid | 0.116 | 63 x 95 x 22 |
| RNA stabiliser port | fluid | 0.010 | 5 x 8 x 5 |
| waste bag (empty at launch, 315 mL capacity), under the reagent bags | fluid | 0.063 | 197 x 95 x 5 |
| waste bag port | fluid | 0.007 | 5 x 8 x 5 |
| -80 C passive cryo cassette: vacuum-insulated body | sample | 0.510 | 78 x 99 x 78 |
| cryo cassette cap | sample | 0.090 | 80 x 20 x 80 |
| microscope base: optics, filter cubes, camera | imaging | 1.137 | 170 x 120 x 70 |
| objective | imaging | 0.073 | 22 x 22 x 16 |
| XY stage frame (travel 96.5 x 57.9 mm: every well over the objective) | imaging | 0.297 | 232 x 151 x 8 |
| stage carriage | imaging | 0.132 | 140 x 98 x 5 |
| X stepper | imaging | 0.116 | 28 x 20 x 20 |
| Y stepper | imaging | 0.116 | 20 x 16 x 20 |
| illumination column | imaging | 0.099 | 16 x 16 x 84 |
| illumination arm | imaging | 0.066 | 112 x 16 x 12 |
| LED head, transmitted light | imaging | 0.165 | 28 x 28 x 22 |
| cassette, 24 wells (SBS footprint, 19.3 mm pitch) | sample | 0.492 | 128 x 86 x 34 |
| culture in the wells | sample | 0.018 | 112 x 73 x 8 |
| cassette lid, sealed and gas-permeable | sample | 0.072 | 128 x 86 x 2 |
| cassette fluid port 1 | sample | 0.009 | 6 x 7 x 6 |
| cassette fluid port 2 | sample | 0.009 | 6 x 7 x 6 |
| fluid manifold | pump | 0.150 | 50 x 94 x 12 |
| micro-pump 1 | pump | 0.125 | 20 x 20 x 26 |
| micro-pump 2 | pump | 0.125 | 20 x 20 x 26 |
| valve 1 | pump | 0.025 | 9 x 9 x 16 |
| valve 2 | pump | 0.025 | 9 x 9 x 16 |
| valve 3 | pump | 0.025 | 9 x 9 x 16 |
| valve 4 | pump | 0.025 | 9 x 9 x 16 |
| controller enclosure (boards, data storage) | electronics | 0.360 | 140 x 93 x 28 |
| connector 1 | electronics | 0.020 | 12 x 8 x 12 |
| connector 2 | electronics | 0.020 | 12 x 8 x 12 |
| bottom panel | housing | 0.061 | 313 x 191 x 2 |
| top panel | housing | 0.061 | 313 x 191 x 2 |
| side panel L | housing | 0.071 | 2 x 191 x 362 |
| side panel R | housing | 0.071 | 2 x 191 x 362 |
| back panel (bay interface) | housing | 0.117 | 310 x 2 x 362 |
| front panel | housing | 0.117 | 310 x 2 x 362 |
| corner post 1 | housing | 0.021 | 8 x 8 x 362 |
| corner post 2 | housing | 0.021 | 8 x 8 x 362 |
| corner post 3 | housing | 0.021 | 8 x 8 x 362 |
| corner post 4 | housing | 0.021 | 8 x 8 x 362 |
| shelf 1 | housing | 0.078 | 310 x 158 x 2 |
| shelf 2 | housing | 0.078 | 310 x 158 x 2 |
| cold wall (to the bay's thermoelectric block) | thermal | 0.028 | 290 x 4 x 258 |
| blind-mate connector: power, data, video | interface | 0.017 | 50 x 10 x 22 |
| grapple fixture base (Dexter-L) | interface | 0.017 | 70 x 4 x 70 |
Measured in orbit
- Cell spreading and shape over time (area, aspect ratio, alignment) Bay brightfield time-lapse, automatic segmentation, 0 g bay vs ground; 1 g bay end points
- Actin cytoskeleton organisation (stress fibres, cortical actin, filopodia/lamellipodia) Lifeact-GFP channel on the bay fluorescence microscope; 40x central-lab images after fixation
- Nuclear size and shape H2B-mCherry channel, live and after fixation
- Actual g-level at each sample Bay accelerometers (quasi-steady about 1e-6 g; arm moves 1e-4 to 1e-3 g) and the centrifuge rotor speed and accelerometer
- Temperature, pressure, humidity and seal state Standard bay sensors, logged with every image
- Radiation dose Active silicon dosimeter plus passive TLDs in each bay
Measured after the return
- YAP/TAZ nuclear-to-cytoplasmic ratio Immunofluorescence on PFA-fixed wells, confocal imaging
- Focal adhesion number, size and composition Immunostaining for vinculin, paxillin and phospho-FAK, with phalloidin for F-actin
- Nuclear envelope and lamina Lamin A/C immunostaining and nuclear morphometry
- Whole-transcriptome changes (adhesion, cytoskeleton, oxidative stress, YAP/TAZ targets such as CTGF and CYR61) RNA-seq on stabilised samples, 0 g vs 1 g in flight vs ground
- Secreted inflammatory factors (for example IL-1 alpha and IL-1 beta, as seen in SPHINX) Multiplex immunoassay on returned conditioned medium
Ground control
Two controls run together. (1) In-flight 1 g control: a second large bay with a small centrifuge holds the same cells at 1.0 g for the whole run. This control shares launch, radiation, the sealed bay and the timeline, so the 0 g vs 1 g difference is the effect of gravity (the approach used in the KUBIK T-cell study, where 1 g samples sat in the central centrifuge positions). (2) Ground control: flight-like 0 g and centrifuge bays in a lab at 37 C, with the same cell lot, media and fluidics script. They replay the flight timeline from telemetry a few hours behind, including real temperature changes and any missed steps. Clinostat or random positioning runs can be added, but they do not replace flight: in the osteocyte study, rotating-wall-vessel runs did not adequately reproduce the flight changes.
Coming home
Our planning values, not a source claim. PFA-fixed wells held in PBS at 4 C are the robust part and can wait up to about 2 weeks. The RNA-stabilised wells are the limit: aim to reach the lab within about 7 days of stabilisation on day 7. With de-orbit on day 9 and 1-2 days of courier time, that is about 3-4 days. The 4 C hold of the chosen stabiliser must be validated in the ground runs. Do not freeze the fixed glass-bottom cassettes. Keep the conditioned medium at 4 C and assay it soon after arrival.
Limits of this flight
- Cells launch frozen and attach in orbit, so the experiment cannot catch the seconds-scale actin response at the onset of microgravity seen on TEXUS rockets. A 1 g to 0 g step on the rotor could, but only with live imaging on the centrifuge.
- Attaching and spreading after an in-orbit thaw is itself a variable. The 1 g centrifuge bay controls for it, but results describe cells that attached in microgravity, not cells adapted from 1 g.
- A small-radius centrifuge is not a perfect 1 g: g changes with radius through the medium height, and there are Coriolis forces and spin-up/stop transients. These must be reported with the data.
- Robotic arm moves give 1e-4 to 1e-3 g. That is equal to or worse than the 1e-4 to 1e-5 g reported on TEXUS-54 and 100-1000 times above the quiet bay level, so arm moves must stay out of key windows and be logged.
- 2 to 14 days covers early and mid-term responses only. Slower processes such as osteocyte maturation in a mineralised 3D matrix over weeks or months need longer missions.
- The passive -80 C cassettes must hold temperature from late loading until thaw on day 1. Their 0.6 kg mass is our estimate and must be checked against the real hold time.
- No crew: there is no manual rescue of a blocked line, bubble or contaminated well. The design relies on extra wells and automated checks.
- The bay microscope (brightfield + 2 fluorescence channels) cannot resolve single focal adhesions live. Adhesion and YAP/TAZ readouts are post-flight end points on fixed cells, and 40x central-lab images are taken only after fixation.
- No CO2 supply: cells must tolerate HEPES-buffered medium in a sealed bay, which can shift pH and metabolism compared with a standard incubator. The ground control must use the same medium.
- Only fixed and stabilised samples come back in good shape. Live cells for functional tests after flight are not part of this protocol, because the return takes days and re-entry and splashdown loads would add artefacts.
A satellite of its own: SP-MECH B · samples return, hardware burns up
SP-MECH is a 140 kg free-flyer. Only the samples come home: 2.5 kg ride in a 45 kg capsule in a drawer behind an optical window; the microscope looks through the window from the service module, and fluid lines cross through dry-break couplings that part at separation once the last fluid step is done. The microscope, pumps and reservoirs, the centrifuge drive stay in the service module and burn up with it, and a small mechanism moves the fixed cassette off the rotor into the capsule first. Why B: both cassettes are fixed before separation; the ring cassette is moved off the rotor into the capsule, which LELP-1's arm does on the lab.
Comes home · capsule
| 0 g glass-bottom cassette | 0.60 |
| 1 g ring cassette (fixed while spinning) | 0.80 |
| sample drawer, optical window, dry-break couplings | 0.50 |
| drawer thermoelectric block | 0.60 |
| Capsule at entry (with heat shield, parachute, hold battery, margin) | 45.0 |
Stays · service module
| experiment instruments, pumps and reservoirs | 5.20 |
| centrifuge drive and structure | 4.50 |
| cassette transfer mechanism (rotor to capsule) | 1.00 |
| Service module dry, with power, radios, attitude, propulsion | 82.4 |
Budgets · estimate
| Launch mass | 139.5 kg |
| Propellant for 190 m/s | 12.0 kg |
| Average / peak load | 88 / 151 W |
| Array, worst eclipse season | 191 W |
| Battery (sized by ascent) | 350 Wh |
| Deorbit burn | 3 × 22 N, 401 s |
Tailored for this experiment
- An on-board 1 g reference centrifuge: this is not a standard bay service. Smallest option (our estimate): a second large bay with a balanced mini-centrifuge, 8 cm radius at about 106 rpm for 1.0 g, carrying a 24-well ring cassette with all wells at the same radius and well floors facing outward. About 5.3 kg (motor and rotor 1.5 kg, ring cassette 0.8 kg, -80 C cryo cassette with cells 0.6 kg, fluidics with reservoirs and rotary coupling 1.2 kg, controller 0.3 kg, structure and containment 0.9 kg) and about 14 W peak. That fits the 6.0 kg / 30 W large-bay limit.
- Do not lay plate-format 12-well cassettes flat on an 8 cm rotor: their wells would sit at roughly 5.4 to 10.6 cm radius, about 0.7 to 1.3 g across the cassette (our calculation). The ring layout keeps g on the cell layer within about 1%.
- Fixation and stabilisation must work on the spinning rotor (rotating fluid coupling or reservoirs on the rotor), so the 1 g samples are never stopped before preservation
- Vibration isolation and dynamic balancing of the centrifuge bay, placement far from the 0 g bay, and accelerometer logging to show the 0 g bay stays near 1e-6 g. Attitude control must absorb the rotor's spin-up momentum (small, about 0.07 N m s for a 1 kg rotor at 8 cm, our estimate), or the rotor needs a counter-rotating mass.
- The 40x central-lab microscope must accept the 1 g ring cassette after fixation, so both arms get the same high-resolution images
Bay or satellite?
The capsule is the floor: no capsule much smaller than about 45 kg has been recovered yet, so 2.5 kg of samples need 140 kg of satellite. On LELP-1 the same samples share one capsule, the whole lab, with 31 other bays. A dedicated satellite earns its place when the experiment needs the schedule to itself, a quieter ride (about 7.3 × 10⁻⁸ g quasi-steady, no robotic arm), or power a bay cannot give.
First-order sizing with a 20 % margin by one model for every experiment, reviewed independently; caveats and sources. The smallest recovered capsule so far (Varda W-1) is under 90 kg, so treat these masses as optimistic.
References
- Versari S, Longinotti G, Barenghi L, Maier JA, Bradamante S. 2013. The challenging environment on board the International Space Station affects endothelial cell function by triggering oxidative stress through thioredoxin interacting protein overexpression: the ESA-SPHINX experiment. FASEB Journal 27(11):4466-4475 https://doi.org/10.1096/fj.13-229195
- Corydon TJ, Kopp S, Wehland M, et al., Grimm D. 2016. Alterations of the cytoskeleton in human cells in space proved by life-cell imaging. Scientific Reports 6:20043 https://pmc.ncbi.nlm.nih.gov/articles/PMC4730242/
- SSC (Swedish Space Corporation). TEXUS 51 and 52 campaign page https://sscspace.com/texus-51-52
- Thiel CS, Tauber S, Lauber B, Polzer J, Seebacher C, Uhl R, Neelam S, Zhang Y, Levine H, Ullrich O. 2019. Rapid Morphological and Cytoskeletal Response to Microgravity in Human Primary Macrophages. International Journal of Molecular Sciences 20(10):2402 https://doi.org/10.3390/ijms20102402
- SSC (Swedish Space Corporation). TEXUS 54/55 successfully launched from Esrange Space Center https://sscspace.com/texus-54-55-successfully-launched-from-esrange-space-center
- Chang TT, Walther I, Li CF, Boonyaratanakornkit J, Galleri G, Meloni MA, Pippia P, Cogoli A, Hughes-Fulford M. 2012. The Rel/NF-kB pathway and transcription of immediate early genes in T cell activation are inhibited by microgravity. Journal of Leukocyte Biology 92(6):1133-1145 https://pmc.ncbi.nlm.nih.gov/articles/PMC3501893/
- Uda Y, Spatz JM, Hussein A, et al., Divieti Pajevic P. 2021. Global transcriptomic analysis of a murine osteocytic cell line subjected to spaceflight. FASEB Journal 35(5):e21578 https://doi.org/10.1096/fj.202100059R
- NASA. Falcon 9 Roars Skyward to Send Dragon on Supply Run to Station (CRS-6), April 2015 https://www.nasa.gov/missions/station/falcon-9-roars-skyward-to-send-dragon-on-supply-run-to-station/
Fact-check: 18 claims and sources checked, 10 confirmed, 8 corrected, 0 removed · confidence high