Immunohistochemistry Techniques, Strengths, Limitations and Applications
Immunohistochemistry Techniques, Strengths, Limitations and Applications
Complete the form below and we will email you a PDF version of "Immunohistochemistry Techniques, Strengths, Limitations and Applications"
Immunohistochemistry (IHC) is an important evaluation tool that falls under the umbrella of immunostaining techniques and exploits antigen–antibody binding to study the status of target molecules in tissues of interest. IHC, in combination with microscopy and image analysis techniques, has become a very powerful tool that offers a direct visualization of tissue antigens using labeled antibodies specific to the antigen. Owing to the versatility, simplicity and affordability of this technique, IHC is now indispensable to the fields of histology, pathology, cancer biology, neuroscience and drug discovery.
Antibodies are small protein molecules that are naturally expressed by the immune system of the body in response to the entry of a foreign molecule (antigen) and help in neutralizing it. Thus, antibodies act as a defense against potentially harmful foreign organisms and their products. The beauty of the antigen–antibody reaction is that each antibody is specific to only a portion of the antigen, called an epitope, and does not bind other molecules that do not match its target, including the body’s own molecules. All immunostaining techniques, including IHC, utilize this important property of the antigen–antibody reaction specificity to ensure the detection of a single molecule type from a milieu of thousands of different ones(Figure 1).
Figure 1: Diagrammatic representation of the specificity of the antigen–antibody reaction that enables detection and localization of a single target in a milieu of thousands of intracellular molecules.
While the development of this important technique could not have been possible without the discovery of antibodies by Emil von Behring and Shibasabura Kitasato in 1890,1 it wasn’t until 1923 that the antigen–antibody complex was detected by Michael Heidelberger using labeled antigens.2 This was followed by John Richardson Marrack’s work describing the nature of the antigen–antibody reaction. Marrack was the first to attach a dye to an antibody in 1934.3,4 The labeling of antibodies with the fluorescent tag, fluorescein, and the detection of their respective antigens in cells and tissues was pioneered by Albert Hewett Coons and others in 19415 and kickstarted the immunostaining revolution.6 The technique has developed and improved much since then and has become a vital tool for discovery and diagnostics.
What is immunohistochemistry (IHC)?
IHC employs a combination of histology, anatomy, immunology and biochemistry to detect the amount, distribution and localization of a specific target within a tissue. The antibodies against the molecule of interest, often a protein, are generated in an organism of a different species and are typically labeled or are aided by another set of labeled antibodies. The tissue samples are prepared following specialized techniques to enable the entry of the antibodies, and the label is detected using light or electron microscopy (Figure 2).
Figure 2: Immunohistochemical staining for tryptophan hydroxylase 2, revealing serotonergic neurons and their projections in human dorsal raphe. Credit: Human Protein Atlas, image available from v21.1.proteinatlas.org.
What’s the difference between immunostaining vs immunohistochemical staining (IHC staining)?
Immunostaining is an umbrella term that encompasses all the techniques that are used for the detection of molecules employing the antigen–antibody reaction. Immunohistochemistry or immunohistochemical staining is a specific use case of immunostaining when the antigen–antibody reaction is used to study the status of molecules in tissue (from the Greek histos, which means tissue).
What’s the difference between immunohistochemistry vs immunofluorescence?
The terms immunohistochemistry and immunofluorescence are applied based on the sample type and method of detection used in the immunostaining technique. Immunohistochemistry refers to the evaluation of target antigens in tissues when the detection method can be either chromogenic or fluorogenic. Immunofluorescence, on the other hand, may refer to both the evaluation of target antigens in cells and tissues and specifically involves the detection of a fluorescent label (fluorophore) (Figure 3).
Figure 3: Immunofluorescence for tyrosine hydroxylase revealing dopaminergic neurons in the mouse midbrain. Credit: Human Protein Atlas, image available from v21.1.proteinatlas.org.
What’s the difference between immunocytochemistry vs immunohistochemistry?
While immunohistochemistry involves the study of the status of antigens in tissue samples, immunocytochemistry refers to immunostaining when the sample of interest is cells in culture (from the Greek cyto, which means cells).
How does immunohistochemistry work?
The success of immunohistochemistry depends on multiple parameters. Therefore, a scientific approach and careful experimental design is key to obtaining reliable and reproducible data.
To give a brief summary of the steps involved,7 IHC is performed on thin slices of tissue obtained from the organism under study. These tissue sections are then processed to permit the entry of antibodies that will bind specifically to the antigen of interest. Routinely, a second antibody will then be applied that binds specifically to the primary antibody and enables detection. In chromogenic detection methods, the secondary antibody is tagged with an enzyme that catalyzes a chromogenic reaction. In immunofluorescence detection methods, the secondary antibody is labeled with a fluorophore that can be directly observed under a fluorescence microscope.
Let’s now consider the different parameters that need to be accounted for when performing IHC.
Tissue: The source of the tissue, i.e., the species of the organism of study and the organ of interest, will determine how the tissue is harvested and prepared for slicing. Prior to performing the experiments, it is vital to refer to literature pertaining to the specific organism to understand the techniques involved in harvesting the tissue and the special care that must be taken when performing IHC. Further, it is important to note that prior to fixation, all tissue must be handled in cold conditions and quickly to prevent rapid decay and drying.
Target: The levels and the subcellular localization of the target are extremely important in IHC experiment design. For example, a more abundantly expressed protein can be detected with less effort and a primary antibody tagged with a fluorophore can suffice for its detection. However, the detection of less abundant targets would require signal amplification methods. The localization of the target within the cell directly influences the degree of permeabilization required. Therefore, while a nuclear protein would require harsher surfactant-based permeabilization treatment, a cytoplasmic component may be detected using a comparatively milder treatment. Detection of intracellular membrane proteins may be achieved by freeze-thawing alone.8
Epitope: The epitope is the small three-dimensional surface region of the antigen to which an antibody would specifically bind. It is important that the epitope to which the antibody would bind is exposed at the time of antibody addition during IHC. The epitope that may be recognized by the antibody can sometimes become masked during tissue processing and steps, such as antigen retrieval, may be necessary to expose the epitope.
Fixation methods: Fixation refers to the preservation of the tissue morphology and cell structure in a stationary state by immobilization of the target. It prevents tissue degeneration and enables long-term storage. Typically, soon after harvesting, the tissue is immersed in an appropriate fixative for several hours before it is further processed for sectioning. To minimize the time between tissue harvesting and fixation and to achieve uniform fixation, whole animal transcardial perfusion with the fixative is the preferred method of fixation in animal models, such as rodents.
Fixatives: The choice of fixative and duration of fixation depends on the tissue and the antigen of interest and requires optimization.9,10 While insufficient fixation might lead to damaged tissue morphology, prolonged fixation can lead to masking of antigens. Fixatives fall into three categories: aldehydes (formaldehyde and glutaraldehyde), alcohols (methanol and ethanol) and acetone-based fixatives. The most commonly used fixative is a 4% (w/v) paraformaldehyde solution prepared in phosphate buffered saline (PBS).
Sample preparation methods: Sample preparation involves embedding the fixed tissue chunk in a matrix that will enable the tissue to be sliced into sections of even thickness. This step needs to be performed carefully to ensure that the tissue is properly aligned in the matrix of choice. Two common methods of sample preparation are formalin-fixed paraffin embedding (FFPE) and freezing. FFPE involves the dehydration of a tissue sample before gradually embedding it in paraffin wax. Freezing involves cryopreservation of the sample using sucrose before embedding it in an optimal cutting temperature (OCT) compound. Once the tissue pieces are paraffin embedded or frozen, they can be stored under appropriate conditions for longer durations.
Sectioning methods: IHC is performed on thin sections of the tissue. The thickness and uniformity of the sections is very important to ensure efficient penetration of the antibody and proper imaging of the tissue. FFPE tissue is commonly sectioned under a microtome at room temperature, whereas frozen tissue is sectioned under a cryostat at sub-zero temperatures. Sections may be obtained along the coronal, sagittal or lateral plane of the tissue. The sections are then mounted on positively charged slides that ensure immobilization of the sections throughout the IHC process.
Pre-processing of tissue sections: Once the tissue sections are ready, they can be processed for penetration of the antibody to facilitate the immunogenic reaction. While the FFPE sections need to be deparaffinized using an organic solvent, such as xylene, followed by a rehydration step, OCT compound embedded sections can be used directly for downstream processing. As frozen sections undergo fewer processing steps, this method is more sensitive for the detection of proteins. However, FFPE affords good morphological preservation and helps achieve thinner sections (as thin as 2 microns).
Antigen retrieval methods: Formaldehyde-based fixation can often lead to masking of the antigen epitopes. Antigen retrieval is therefore required to unmask the epitopes and make them available for antibody binding and is often performed.11 This step is important when performing IHC for FFPE tissue, but can be too harsh for frozen tissue sections and may be omitted. Antigen retrieval can be achieved either by the application of heat (heat induced epitope retrieval: HIER) or through enzymatic degradation (proteolytic-induced epitope retrieval: PIER) in an appropriate buffer.
Permeabilization: Permeabilization is an essential first step that renders the plasma membrane of the cells in the tissue porous, thus allowing the entry of IHC reagents and antibodies. Routinely, surfactants such as Triton X-100, Tween-20, saponin and digitonin are used to achieve permeabilization. However, for a gentler permeabilization for preservation of intracellular membranes, the sections may be subjected to the freeze-thaw process. Fixatives, such as methanol and acetone, also permeabilize the tissue, and when using these fixatives this step may be omitted. The concentration of the surfactant and the time of incubation are determined based on factors such as the fixative used, thickness of the tissue section and the subcellular localization of the antigen of interest.
Blocking buffer: Although antibody–antigen binding can be very specific, some antibodies may adhere to certain non-specific cellular components due to various intramolecular forces at play. Incubation in blocking buffer before addition of the antibody helps to prevent the non-specific binding of the antibodies in the tissue. Commonly used blocking agents are normal serum and bovine serum albumin. When using chromogenic detection, blocking of endogenous enzyme activity is also required and can be achieved by using hydrogen peroxide to block endogenous peroxidase activity or levamisole to block endogenous alkaline phosphatase activity.
Detection method: The detection of the target antigen may be direct, where the label is directly attached to the primary antibody, or indirect, where the label or an enzyme that catalyzes a chromogenic reaction is attached to a secondary antibody. Further, signal amplification methods may be applied to enhance the sensitivity of signal detection.
Primary antibodies: The antibody that directly binds to the epitope of the antigen of interest is called the primary antibody. Primary antibody selection is an important step in IHC experimental design12 and one needs to ensure that the selected antibody is specific to the species under study. Specificity of the antibody against the target antigen also needs to be thoroughly evaluated. Primary antibodies may be polyclonal or monoclonal: while polyclonal antibodies consist of multiple individual antibody molecules that can recognize different epitopes of the same target, monoclonal antibodies all recognize the same single epitope. The concentration of the primary antibody needs to be determined through careful experimentation to achieve the best results.
Secondary antibodies: The secondary antibodies are antibodies that recognize the primary antibody and thus enable detection of the target antigen. Secondary antibodies are often tagged with an enzyme to facilitate signal amplification for chromogenic detection, or they may be tagged with a fluorophore.
Signal amplification: For target antigens that are expressed at low levels, signal amplification may need to be performed to improve the sensitivity of the technique. Several signal amplification strategies that are used include, the avidin-biotin complex (ABC) method, labeled streptavidin-biotin (LSAB) method and tyramide signal amplification (TSA).13
Label: Labels used to detect the target antibody are attached to the primary or secondary antibodies and may be either fluorogenic (as in the case of immunofluorescence) or chromogenic. Fluorogenic labels, such as fluorescein and tetramethylrhodamine (TAMRA), can be directly viewed under a fluorescence microscope. Chromogenic methods involve the conversion of a chromogenic substrate, such as 3,3’-diaminobenzidine (DAB) and 5-bromo-4-chloro-3-indolyl-phosphate (BCIP)/nitro blue tetrazolium (NBT), to a colored product in the presence of an antibody-conjugated enzyme such as, horseradish peroxidase (HRP) or alkaline phosphatase (AP).
Counterstain: The tissue is often stained with a secondary nuclear or cytoplasmic stain that labels all the cells and helps visualize the IHC-labeled cells against the general morphology of the tissue, thus providing a contrast. Counterstains used with chromogenic IHC labels include hematoxylin, nuclear fast red and eosin; while those used with fluorogenic IHC labels include 4′,6-diamidino-2-phenylindole (DAPI), Hoechst33342 and propidium iodide.
Mounting: The prepared tissue sections are mounted in a medium with an appropriate refractive index that facilitates imaging under a microscope, protects the fluorescently labeled sections from photobleaching and prevents the section from drying. Commonly used mounting media include DPX, synthetic resins and glycerol-based mounting media containing antifade agents for fluorescently labeled sections.
Multiplexing: Multiplexing is performed when an experiment requires the concomitant investigation of more than one target antigen. Fluorophore labeling is the preferred method for multiplexing due to the availability of tags that fluoresce in the entire range of the visible spectrum and adjacent wavelengths. Special care must be taken to prevent the cross-reactivity of secondary antibodies when performing multiplexing.
Imaging method: Chromogenic labels can be detected using a light microscope. Fluorescence or confocal microscopes are used for the detection of fluorophores. Electron microscopy may be used for imaging after immunohistochemical labeling with colloidal gold particles.14
Controls: Proper controls are indispensable to ascertain the authenticity of the results from an IHC experiment. A tissue sample in which the expression of the target antigen is known can serve as a positive control. Similarly, a sample wherein the target antigen is known to be absent can be used as a negative control. Additionally, antibody controls need to be used to verify the specificity of the antibody.12
An example immunohistochemistry protocol
Putting all these variables together, a generalized protocol for IHC would include the following steps. 15 (Figure 4).
- Fixation: Target tissue is harvested from the organism and fixed in 4% (w/v) formaldehyde for 16–24 h at 4 °C.
- Sample preparation and sectioning: For FFPE sections, tissue is dehydrated using an alcohol gradient, paraffin embedded and 2–4 mm thick sections are obtained and mounted on charged slides. For frozen sections, tissue is embedded in OCT compound and 4–8 mm thick sections are sliced and mounted on charged glass slides.
- Deparaffinization: This step is performed only for FFPE sections. The sections on glass slides are deparaffinized in xylene and then rehydrated using an alcohol gradient.
- Antigen retrieval: HIER is generally preferred and can be performed by incubating the sections in citrate buffer (pH 6.0) in a pressure cooker (3 min under pressure) or a microwave oven (20 min). Sections are then washed with PBS Tween-20 twice (2 x 5 min).
- Peroxidase blocking: To block endogenous HRP activity, sections are incubated in 3 % (v/v) hydrogen peroxide solution for 15 min at room temperature. This is followed by three PBS Tween-20 washes (3 x 5 min).
- Permeabilization: This step is optional; sections may be incubated in 0.1 % (v/v) Triton X-100 in PBS for 10 min at room temperature. Sections are washed with PBS Tween-20 thrice (3 x 5 min) before proceeding to the next step.
- Blocking: Sections are incubated with 3–5% (v/v) of normal serum of the same species in which the secondary antibody was produced for 30 min at room temperature.
- Primary antibody: Sections are incubated in species-matched primary antibody at the optimized concentration for 1–4 h at room temperature or overnight in a hydration chamber at 4 °C. The primary antibody can be diluted in PBS or in blocking buffer. This is followed by two PBS Tween-20 washes (2 x 5 min).
- Secondary antibody: Sections are then incubated in biotinylated secondary antibody diluted to an appropriate concentration (e.g., 1:1000) in PBS for 30 min–1 h at room temperature. Sections are washed thrice with PBS Tween-20 (3 x 5 min).
- Signal amplification: This is followed by incubation of the sections in streptavidin-HRP in PBS for 30 min at room temperature. Sections are rinsed thrice with PBS Tween-20 (3 x 5 min).
- Chromogenic detection: Sections are then incubated with solution containing 1% (w/v) DAB (250 μL) and 0.3% (v/v) hydrogen peroxide (250 μL) in 5 ml of PBS for 1–3 min at room temperature until a brown color has developed. This is followed by three washes with distilled water (3 x 5 min).
- Counterstain: The sections can be counterstained by incubation in hematoxylin for 1 min.
- Mounting: Sections are then dehydrated using an alcohol gradient and mounted in DPX.
- Imaging: The sections can be imaged using a light microscope.
Figure 4: Diagrammatic representation of the key steps involved in IHC.
Strengths and limitations of immunohistochemical analysis
Some of the main strengths and limitations of IHC are summarized in Table 1.
Table 1: Strengths and limitations of IHC.
Affordable and simple procedure that can be performed with few resources
Specificity of antibodies can be variable and needs to be thoroughly checked using appropriate controls
Powerful technique to study localization and presence/absence of a target at the tissue and cellular level
The method is semi-quantitative, and the absolute abundance of the target cannot be reliably determined
Paraffin embedded and frozen tissue samples can be stored and accessed when required
Tissue is highly processed and may lead to loss of information of the natural state
Stained tissue sections can be stored and referred to whenever required
IHC is a multi-step procedure and variability can be introduced at any stage leading to poor reproducibility of results16
What is immunohistochemistry used for?
Immunohistochemistry is a simple and cost-effective technique that has become an important tool for pathologists and scientists. Some applications of this technique are listed here:
- Biomarker assessment in oncology: IHC is a very popular tool for detection of biomarkers in cancer diagnosis as well as for the development of new biomarkers. IHC allows tumor detection, staging and classification, in addition to predicting tumor prognosis and understanding the response of tumor to treatment paradigms. IHC-based biomarkers have become vital for the diagnosis and treatment of breast cancer,17 prostate cancer,18 pancreatic cancer,19 lung cancer,20 bladder cancer,21 colorectal cancer22 and ovarian cancer.23
- Diagnoses of infectious diseases: IHC serves as an important tool for detection and identification of pathogenic antigens in tissue samples from infected individuals, which can help in the treatment of infectious diseases.24 Bacterial pathogens such as, Bartonella quintana, Yersinia pestis, Treponema pallidum, Chlamydia trachomatis; viral pathogens such as, human herpesvirus type 8 (HHV8), Epstein-Barr virus (EBV), human immunodeficiency virus (HIV); fungal pathogens such as, Candida albicans and Cryptococcus neoformans var. gattii; and protozoal pathogens such as, Plasmodium falciparum and Trypanosoma cruzi have been successfully identified using IHC.
- Evaluating neurodegenerative disorders: Abnormal protein conformations and aggregations can be identified and evaluated using IHC and are a routine feature of many neurodegenerative disorders including, Alzheimer’s disease (AD), chronic traumatic encephalopathy (CTE), progressive supranuclear palsy (PSP), Pick’s disease, Lewy body disease, multiple system atrophy (MSA) and amyotrophic lateral sclerosis (ALS).25
- Human Protein Atlas: IHC has contributed to the age of big data by enabling the mapping of the human proteome.26 The Human Protein Atlas27 is a freely available and valuable resource that contains tissue level information of expression encompassing 90% of all protein encoding genes.
1. Serum Therapy in Therapeutics and Medical Science. The Nobel Prize. https://www.nobelprize.org/prizes/medicine/1901/behring/lecture/. Published December 12, 1901. Accessed June 24, 2022.
2. Heidelberger M, Avery OT. The soluble specific substance of pneumococcus: second paper. J Exp Med. 1924; 40(3):301-317. doi:10.1084/jem.40.3.301
3. Marrack JR. Nature of antibodies. Nature. 1934;133: 292–293. doi: 10.1038/133292b0
4. Marrack JR. The chemistry of antigens and antibodies. J Phys Chem. 1934; 38(7): 989-989. doi:10.1021/j150358a015
5. Coons AH, Creech HJ, Jones RN, Berliner E. The demonstration of pneumococcal antigen in tissues by the use of fluorescent antibody. J Immunol. 1942; 45(3): 159-170. https://www.jimmunol.org/content/45/3/159
6. Childs GV. History of immunohistochemistry. In: McManus L, Mitchell R eds. Pathobiology of Human Disease: A Dynamic Encyclopedia of Disease Mechanisms. 1st ed. San Diego: Academy Press. 2014: 3775-3796. doi: 10.1016/B978-0-12-386456-7.07401-3
7. Magaki S, Hojat SA, Wei B, So A, Yong WH. An introduction to the performance of immunohistochemistry. Methods Mol Biol. 2019;1897:289-298. doi:10.1007/978-1-4939-8935-5_25
8. Mardones G, González A. Selective plasma membrane permeabilization by freeze-thawing and immunofluorescence epitope access to determine the topology of intracellular membrane proteins. J Immunol Methods. 2003;275(1-2):169-177. doi:10.1016/s0022-1759(03)00015-2
9. Burns JA, Li Y, Cheney CA, et al. Choice of fixative is crucial to successful immunohistochemical detection of phosphoproteins in paraffin-embedded tumor tissues. J Histochem Cytochem. 2009;57(3):257-264. doi:10.1369/jhc.2008.952911
10. Paavilainen L, Edvinsson A, Asplund A, et al. The impact of tissue fixatives on morphology and antibody-based protein profiling in tissues and cells. J Histochem Cytochem. 2010;58(3):237-246. doi:10.1369/jhc.2009.954321
11. Pileri SA, Roncador G, Ceccarelli C, et al. Antigen retrieval techniques in immunohistochemistry: comparison of different methods. J Pathol. 1997;183(1):116-123. doi:10.1002/(SICI)1096-9896(199709)183:1<116::AID-PATH1087>3.0.CO;2-2
12. Ivell R, Teerds K, Hoffman GE. Proper application of antibodies for immunohistochemical detection: antibody crimes and how to prevent them. Endocrinology. 2014;155(3):676-687. doi:10.1210/en.2013-1971
13. Dunkenberger L, Del Valle L. Antigen retrieval and signal amplification. Methods Mol Biol. 2022;2422:65-74. doi:10.1007/978-1-0716-1948-3_5
14. Faulk WP, Taylor GM. An immunocolloid method for the electron microscope. Immunochemistry. 1971;8(11):1081-1083. doi:10.1016/0019-2791(71)90496-4
15. Kim SW, Roh J, Park CS. Immunohistochemistry for pathologists: Protocols, pitfalls, and tips. J Pathol Transl Med. 2016;50(6):411-418. doi:10.4132/jptm.2016.08.08
16. O'Hurley G, Sjöstedt E, Rahman A, et al. Garbage in, garbage out: a critical evaluation of strategies used for validation of immunohistochemical biomarkers. Mol Oncol. 2014;8(4):783-798. doi:10.1016/j.molonc.2014.03.008
17. Zaha DC. Significance of immunohistochemistry in breast cancer. World J Clin Oncol. 2014;5(3):382-392. doi:10.5306/wjco.v5.i3.382
18. Carneiro A, Barbosa ÁRG, Takemura LS, et al. The role of immunohistochemical analysis as a tool for the diagnosis, prognostic evaluation and treatment of prostate cancer: A systematic review of the literature. Front Oncol. 2018;8:377. doi:10.3389/fonc.2018.00377
19. Luu TT. Review of immunohistochemistry biomarkers in pancreatic cancer diagnosis. Front Oncol. 2021;11:799025. doi:10.3389/fonc.2021.799025
20. Inamura K. Update on immunohistochemistry for the diagnosis of lung cancer. Cancers (Basel). 2018;10(3):72. doi:10.3390/cancers10030072
21. Wu J, Wen JM, Wang YC, et al. Prognostic value of an immunohistochemical signature in patients with bladder cancer undergoing radical cystectomy. Front Oncol. 2021;11:641385. doi:10.3389/fonc.2021.641385
22. Banias L, Jung I, Bara T, et al. Immunohistochemical-based molecular subtyping of colorectal carcinoma using maspin and markers of epithelial-mesenchymal transition. Oncol Lett. 2020;19(2):1487-1495. doi:10.3892/ol.2019.11228
23. Köbel M, Rahimi K, Rambau PF, et al. An immunohistochemical algorithm for ovarian carcinoma typing. Int J Gynecol Pathol. 2016;35(5):430-441. doi:10.1097/PGP.0000000000000274
24. Oumarou Hama H, Aboudharam G, Barbieri R, Lepidi H, Drancourt M. Immunohistochemical diagnosis of human infectious diseases: a review. Diagn Pathol. 2022;17(1):17. doi:10.1186/s13000-022-01197-5
25. Dugger BN, Dickson DW. Pathology of neurodegenerative diseases. Cold Spring Harb Perspect Biol. 2017;9(7):a028035. doi:10.1101/cshperspect.a028035
26. Sjöstedt E, Zhong W, Fagerberg L, et al. An atlas of the protein-coding genes in the human, pig, and mouse brain. Science. 2020;367(6482):eaay5947. doi:10.1126/science.aay5947
27. Uhlén M, Fagerberg L, Hallström BM, et al. Proteomics. Tissue-based map of the human proteome. Science. 2015;347(6220):1260419. doi:10.1126/science.1260419