Interferons are pleotropic cytokines with antiviral, antiproliferative and immunomodulatory activities. Activity can be assayed by any of these methods; however the standard assay is protection of cells from the cytopathic effect (CPE) of certain viruses. In part, this preference is historical, since this was the initial activity which defined these molecules. The other reason that the CPE assay is used to characterize the activity of interferons (IFN) is that it is amongst the most sensitive activity assay, with often >100 fold more sensitivity than the alternatives. In fact, given that the Kd of IFN on human cells is ~0.5 nM and in some antiviral assays the IFN is effective at 0.5 pM, very little receptor occupancy is required for protection of the test cells from viral challenge.

Many combinations of cells and viruses can be used, but certain pairs tend to be used most frequently. To measure mouse interferons, researchers usually use L929 cells with either vesicular stomatitis virus (VSV), often the New Jersey strain, or encephalomyocarditis virus (EMCV). This cell line is sensitive to alpha, beta and gamma interferon. For the measurement of human interferon a wider variety of systems are used. Many researchers will use the human lung carcinoma cell line A549 due to their ease of growth and relatively high sensitivity to all forms of human interferons. These cells are usually challenged with EMCV. Human alpha interferon can be assayed using the bovine kidney line MDBK challenged with VSV. One advantage of this system is the relative insensitivity of the bovine cells to human IFN-beta and gamma giving some selectivity in uncharacterized samples. Human IFN-beta can also be measured on the green monkey kidney cell line Vero challenged with VSV. The advantage of this assay is that human IFN-alpha is much less active on these cells than IFN-beta.

Treatment of cells with IFN leads to the production of a whole host of proteins, many of which interfere with viral replication. As noted above, very little receptor occupancy is required to lead to effective protection, but some time is required for the synthesis of the new proteins. Generally, IFN is pre-incubated with the cells for 4 (rapid assay) to 24 hours (sensitive assay) prior to viral challenge. Depending upon the cell and virus combination, the cells are then incubated for a further 24-48 hours to effect nearly complete killing of the unprotected cells.

CPE assays are generally developed as follows. Cells are plated in 96 well tissue culture plates at defined density. This density is chosen to ensure that cells not challenged with virus are just confluent at the end of the assay. Interferon is titrated by two fold dilution across the plate (12 data points) in duplicate or triplicate. Often a pre-dilution is required. The goal is to dilute the sample sufficiently to have the interferon concentration range from protecting the cells to not protecting roughly near the center wells of the plate. In addition, a standard is included to allow determination of the activity in Units/ml. These standards are generally laboratory standards which have been calibrated by numerous repeated assays using one of the International standards. Also included is one row which receives no IFN. After incubation of the cells with IFN, virus is added to all the wells which received either IFN sample or standard. Virus is also added to 6 wells in the row without IFN, while the other 6 wells receive media with no virus. The 6 wells receiving virus but no IFN serve as the “virus control” and provide the control for 0% protection. The 6 wells receiving no IFN and no virus serve as the “cell control” and provide the control for 100% protection. The amount of virus to be added is determined empirically. Usually, the virus will be added between 0.1 and 1 MOI and approximates the amount of virus that kills very close to 100% of the untreated cells in the desired incubation time.

The plates are monitored microscopically occasionally during the incubation time. Once viral killing is complete, the assay is developed. The classic, and still most common, method involves fixing and staining the plates with an acid/methanol solution containing Crystal Violet dye. This dye will stain any remaining cells attached to the plate. Alternatively vital stains such as neutral red can be used. Assays can also be developed using dyes such as XTT, MTT or MTS which measure mitochondrial activity or surviving cells. In the crystal violet assay, the excess stain is washed off the cells with water or PBS, and the plate is then dried.

Figure 1. A549/EMCV CPE Assay for Alpha IFN

Rows B and C contain an unknown sample. Row F contains the IFN lab standard.

Row G, well 1-6 are the cell control and wells 7-12 are the virus control.

Once dry, the plate is examined microscopically. Wells are examined and the well in which 50% of the cells are alive, the endpoint, is noted. This is also noted for the standard. Since the Unitage of the standard is known, it can be used to calculate the U/ml of the unknown sample. For example, a standard of 2500 U/ml reaches an endpoint at well 4 on the plate. If the unknown sample has not been pre-diluted and reaches and endpoint at well 9, it contains 32 times the protective activity based on fold dilution/well raised to the number of wells different (2⁵ = 32). 2500 U/ml standard endpoint times 32 yields 80,000 U/ml for the concentration of the interferon in the unknown. Sometimes a sample will reach an endpoint between wells. Say there is 80% protection in well 9 and 20% protection in well 10, the endpoint is recorded as well 9.5. The Unitage would then be the geometric mean of the values obtained at wells 9 and 10. In the above example, well 8 would be 80,000 U/ml, well 9 would be 160,000 U/ml and well 9.5 would be 113,137 U/ml. Generally, within a single assay the endpoint will be within 1 well for replicates. The calculated values for each replicate are then arithmetically averaged to obtain the U/ml for the sample. Thus a sample that gave endpoints of 8 and 9 would have unitage of 120,000 (Average of 80,000 and 160,000).

If the concentration of the IFN in the samples is known, it is also then possible to calculate the specific activity. To do this, one divides the U/ml by the concentration in mg/ml, the results are then expressed as U/mg. In the above example, if the concentration of the sample is 0.0004 mg/ml the specific activity would be 3.0 x 10⁸ U/mg (calculated as 120,000 U/ml divided by 0.0004 mg/ml), a value which is typical for an IFN alpha or beta sample.

A further analysis can be done by solubilizing the dye in the cells using a solution such as 70% MeOH. This then allows determination of the absorbance in the wells using a plate reader. Again, using the cell control as 100% protection and the virus control as 0% protection a curve can be plotted and the EC50 determined either in U/ml or pg/ml.

Figure 2. Graphic Analysis of IFN CPE Assay

Data were fit to a sigmoidal dose response curve (variable slope) using the GraphPad Prism software package. The EC₅₀ values were determined for each sample as the dilution from the stock required to give 50% protection. From the known unitage of the Lab Standard it is possible to determine the IFN concentration required to protect 50% of the cells. In this case 1.09 U/ml gives 50% protection. The U/ml is then the Std. U/ml EC₅₀ times this dilution. Again, knowing the concentration of the sample it is possible to determine the pg/ml which gives 50% protection and the Specific Activity of the samples.

Since bioassays are inherently complex assays which vary due to the metabolic state of the cells, the ability of the IFN to protect the cells, and the virus replication, there will be greater variability in these assays than mass based assays such as ELISAs. For example, during the calibration of the international standard for beta interferon which was a multi-site, multi-assay study, the geometric CVs obtained were on the order of 60-80%. Generally in these assays, results from run to run and operator to operator can vary by about 1/2 log from the actual value. Thus, in order to obtain solid numbers, multiple runs are required. With experience and careful control, the CV value can be reduced (related service link here).

Below are the results from 3 different operators over a 4 week period for the Lab Standard in the A549/EMCV assay. The nominal titer (determined by 20 runs versus the international standard) is 4276 U/ml. In this study, the average potency is 3996 and the CV is 33%. The geometric mean is 3885 and the geometric CV is 13%.

In summary, there are multiple methods available to measure the potency of a particular IFN preparation. The standard assay measures protection from viral challenge. However there can be significant variation in the results obtained in different assay systems with a particular interferon.

Measuring interferon-alpha levels.

IFN-α protein levels can be determined by direct methods including directly determining mass levels and indirectly by many unique biological activity assays. There have been many assays developed over the years for the detection of IFN-α. While many were important for increasing our understanding of IFN-α expression and activity, many were very laborious and low-throughput. The assays highlighted here are those of a medium to high throughput in nature that could be adapted to large scale drug screening and/or clinical sample analysis. Both purified IFN-α preparations as well as IFN-α from stimulated cells can be analyzed by the methods described below.

Direct protein analysis

The sandwich Enzyme Linked ImmunoSorbent Assay (ELISA) has become an invaluable tool for the rapid and highly quantitative analysis of cytokines including IFN-α. Most of these commercial assays work in complex sample matrices including tissue culture medium, serum, plasma, cerebrospinal fluid and urine. Advantages of this approach are that these assays are rather simple, rapid and they are highly specific for IFN-α. Limitations are the possibility of false positive results due to the presence of heterophilic binding agents and false negatives due to the interaction of the IFN-α with soluble receptors, antibodies or other factors that would inhibit IFN-α capture and detection. One area that is difficult to control is individual IFN-α subtype detection capacity. Many IFN-α ELISA kits are developed to detect IFN-α2. One must make sure when using an IFN-α ELISA kit that it has been developed to detect most if not all of the IFN-α subtypes. This way, you can have accurate global expression data pertaining to the IFN-α family of proteins and just not IFN-α2.

An alternative to solid-phase ELISA are the solution based bead-based technology systems where the assay sample is mixed with the anti-IFN-α beads and the detection antibody in solution. The sample is then passed through a reader that detects and quantifies the level of bound IFN. Similar to the more complex ELISA systems, these reagents are more costly and require dedicated reader systems. A distinct advantage of these systems over standard ELISA is the potential to detect multiple cytokines simultaneously within a single sample. Another method which is limited to cells in culture involves the detection of cells expressing IFN-α by flow cytometry methods. This is not as sensitive as ELISA and does require a flow cytometer and a skilled operator. However, these assays are useful when performing studies to look for distinct IFN-α expressing cell populations in mixtures.

In summary, the single analyte colorimetric anti-IFN-α continues to be the simplest and most affordable means to monitor IFN-α protein levels. The more expensive and laborious procedures are valid alternatives if the proposals warrant the sensitivity or multiplex capabilities of these approaches.

Biological Activity Assays

The assay of IFN-α biological activity by most methods is prone to variation. Early studies conducted during the clinical development of IFN-α2 showed that significant variation was common between different laboratories and operators^(7,8). Different cell line origins, passage numbers, media constituents and virus preparations were all sources of possible assay variation. Consequently, reference standards were prepared to be analyzed in parallel with test samples to calculate activity levels. These reference standards are available from several sources and are the utmost critical reagent for any assay method chosen. Since these standards are in finite levels, they are only used to calibrate internal lab standards that are run in all assays each and every time and preferably, on each assay plate.

The cytopathic protective effects (CPE) assay continues to be the most widely used assay to determine IFN-α biological activity. Briefly, IFN-α responsive cells are treated with serial dilutions of test samples and know dilutions of lab standards. Cells are then challenged with a single concentration of cytopathic virus. A second incubation step is carried out and completed when cells treated with only virus show complete killing and lysis. IFN-α activity values of the experimental samples are determined by comparing dilution endpoints to those of the lab standards. Advantages of this approach are low reagent costs and high sensitivity. Disadvantages include cross-reactivity with other type I cytokines (and potentially other cytokines depending on the cell line and virus used), extensive training required to perform assays on a reproducible level, multiple addition steps and overall assay times (usually 24 to 72 hrs). Moreover, biological activity can vary greatly in a cell and virus type manner. Therefore, if these experiments are going to be used as screening assays, one must have sufficient stocks of cells and virus to maintain assay continuity over time.

Another clinical application of IFN-α proteins is as an anti-cancer agent. Consequently, IFN-α is also screened for the ability to inhibit the replication of cancer cells in culture and therefore a desired screening assay can be used to measure the antiproliferation capacity of IFN-α. These assays are similar to antiviral assays in that they are carried out by treating cells with experimental sample and lab standard dilutions in parallel and activity is determined by comparison of experimental and lab standard endpoints. The advantages of these assays are they require less reagent addition steps than antiviral assays and have good sensitivity. Disadvantages include an extended 3 to 7 day overall assay time and similar to antiviral assays, the potential for cross-reactivity with other cytokines. Nevertheless, these assays are the standard methods for monitoring the antiproliferative effects of interferon-alpha. Another cell-based assay option that indirectly measures IFN-α levels involves monitoring specific gene expression induced by IFN-α in a dose-dependent manner. Preferably, cell lines containing a stably integrated gene construct that contain a promoter region from an interferon stimulated gene (ISG) controlling translation of a reporter gene are employed. Addition of serial dilutions of experimental and lab standards are conducted in parallel and unknown values are determined by backfitting values against the lab standard curve (similar to ELISA data analysis). The primary advantages of these types of assays are comparable sensitivity to antiviral assays without the multiple reagent addition steps and extended incubations times. These assays are simple and rapid when compared to other cell based assays, while still maintaining the sensitivity. Another advantage is that one can employ the use of cryopreserved cells to limit the need for continuous cell culture which also limits assay variability. Disadvantages are similar to all other bioassays in that interfering substances can artificially dampen or enhance the reporter signal. Nevertheless, these newer assays represent a quicker, simpler and more reproducible means to assess IFN-α biological activity. Lastly, flow cytometry can also be employed for indirect IFN-α gene upregulation. Class I major histocompatibility complex is up-regulated by IFN-α. While more labor intensive and less quantitative, it does allow for differentiating the cell types that respond to IFN-α by sorting populations for other cell markers. Therefore this would be a highly useful research tool to determine if IFN-α expression levels could be traced to specific cell types.

In summation biological assays, although generally more labor intensive than direct protein assays are critical for the proper evaluation of IFN-α biological activity. One cannot assume biological activity based on protein levels. As such, if activity levels are required for your studies, a bioassay must be carried out. Lastly, the appropriate cell line and additional reagents described for the assays aforementioned need to be carefully evaluated for sensitivity, specificity and reproducibility in relation to the origin of your samples and the assay sensitivities required.

Other Options

There are other assays that can be used as well. For example, signal transducer and activator of transcription proteins 1 and 2 (STAT1 and STAT2) are phosphorylated in response to IFN-α treatment and form the core of the transcription unit (along with interferon regulatory factor 9; IRF9) that upregulates ISG expression. These assays can be performed either by ELISA with cell lysates or by high content screening assays. The primary advantages of analyzing STAT1 or STAT2 is that phosphorylation occurs in minutes and assays can be performed within a single day. Drawbacks include the need for cell lysis in the ELISA, and the cost of high content screening equipment. However, the rapidity of these assays makes them a plausible choice especially for screening large compound libraries. In addition, qRT-PCR based detection is possible, but not established for detecting all the IFN-α subtypes. If developed this could be a useful tool for the early and rapid IFN-α production in stimulated cells.

Summary

Are there any perfect IFN-α assays? Clearly, there is no single, absolute assay for measuring IFN-α. This is due in part to the nature that they are a family of proteins and also that they are usually expressed with several other pro-inflammatory cytokines that may also produce overlapping biological effects. Perhaps the best approach is to screen both protein mass and biological activity in parallel. Following a series of initial spike recovery studies, it should be possible to have a strong correlation between the two assays. This way, each assay will indirectly serve as an assay control for the other.

Today, the advancement of immune response modulating drugs into clinical development is demanding better assays to monitor cytokines like IFN-α. In one instance, TLR agonists are being developed to specifically enhance IFN-α expression, while limiting the expression of TNF-α⁽⁹⁾. Likewise, newer cancer therapies are being developed that IFN-α is a critical component. Separately, there is a growing body of evidence that IFN-α may be a causative agent either predisposing or advancing autoimmunity diseases like SLE and that monitoring IFN-α levels may be important for controlling disease progression. A recent search of the NIH clinical trials site www.clinicaltrials.gov using “interferon alpha” as the query returned over 560 hits including recently completed, recruiting and ongoing clinical trials. Monitoring this cytokine will remain an important goal for many areas of therapies for years to come.

Lastly, one higher order question remains, are there really “good” and “bad” IFN-α subtype proteins? Would some IFN-α subtypes be better therapies for antiviral and cancer treatments, producing more potent desired activities with fewer side effects? Conversely, could bad IFN-α subtypes associated with autoimmune diseases be selectively blocked, preventing the disease effects but still allowing ample host response to pathogens? These questions cannot be adequately addressed in large scale studies to date because no reagents are commercially available to differentiate all of the individual IFN-α proteins. Hopefully, current studies will help to better determine global levels of IFN-α in a variety of settings so that this information can be used as a base for deconvolution of the individual IFN-α subtypes when these assays are established.

References:

1. Heathcote EJ, “Antiviral therapy: Chronic Hepatitis C”, Viral Hepatitis, Suppl. 1:82-88, 2007.
2. Pascual V, Farkas L, and Banchereau, “Systemic lupus erythematosus: all roads lead to type I interferons”, Current Opinion in Immunology,18:676-682.
3. Pestka S, Ed., Interferons, Methods in Enzymology, 78: Part A, 1981.
4. Pestka S, Ed., Interferons, Methods in Enzymology, 79: Part B, 1981.
5. Pestka S, Ed., Interferons, Methods in Enzymology, 119: Part C, 1986.
6. Meager A, “Biological assays for interferons”, Journal of Immunological Methods, 261:21-36.
7. Pestka S, Meager A, “Interferon standardization and designations”, Journal of Interferon Cytoine Research, 17: Suppl 1:S9-14
8. Meager A, Gaines Das R, Zoon K, Mire-Sluis A, “Establishment of new and replacement World Health Organization International Biological Standards for human interferon alpha and omega”. J. Immunological Methods. 257 (1-2):17-33, 2001
9. Thomas A, Laxton C, Rodman J, Myangar N, Horscroft N, Parkinson T, “Investigating Toll-like receptor agonists for potential to treat hepatitis C virus infection”, Antimicrobial Agents and Chemotherapy, 51(8):2969-78, 2007.

The family of Type I interferons, which include the Interferon-alpha (IFN-α) proteins and Interferon-beta (IFN-β), are secreted by many cell types including macrophages, lymphocytes, fibroblasts, endothelial cells and others. They stimulate cells of the immune system including both macrophages and NK cells to form an antiviral response, and they have also been shown to be active against tumors. IFN-β is known to also be produced in the central nervous system (CNS) by both the glial cells and astrocytes. Endogenous IFN-β production is thought to be mainly regulated by signaling through toll-like receptors (TLRs) that respond to ligands expressed by pathogens. However, several studies suggest that the intracellular signaling pathways for IFN-β induction are different for viral and non-viral inducers, and each pathway is likely to depend on discrete second messenger mechanisms. Each signaling pathway can induce or modulate a varied subset of other cytokines and chemokines that contribute to the innate immune inflammatory response. All of the Type I interferons have been shown to have important roles in regulating innate and adaptive immune responses.

Multiple sclerosis (MS) is an inflammatory disease of the CNS. In the last decade, there has been an increase in the number of studies trying to determine the immune mechanisms that mediate tissue damage in MS (Javed and Reder, 2005; Weiner, 2008). IFN-β is effective for treating Multiple Sclerosis (MS); however the mechanism by which the therapeutic effect takes place remains somewhat elusive to date (Pozzilli and Prosperini, 2008). It is thought that among other pathogenic targets, IFN-β down-regulates T-cell activation by altering the expression of proteins involved in antigen presentation, and studies indicate that it also promotes the differentiation of activated T cells away from a T-helper-type 1 Th1 response (pro-inflammatory) and towards a Th2 response (anti-inflammatory). Accordingly, IFN-β decreases the expression of IL-12 and promotes the expression of anti-inflammatory cytokines like IL-27 by macrophages. (Baccala et al., 2005).

Systemic administration of IFN-β has been shown to slow MS disease progression, reduce relapse rate, and decrease the number of CNS lesions. Although the signaling pathways remain to be deduced, results clearly implicate the involvement of IFN-β in regulating the inflammatory response in the CNS. Other autoimmune diseases similar to MS that are characterized by a T-lymphocyte delayed-type hypersensitivity response include rheumatoid arthritis (RA) and Type I diabetes mellitus (DM). The diseases are caused by the similar immune deregulation found in the synovial membrane lining the non-cartilaginous surfaces of the joints, and the b islet cells of the pancreas, respectively. FDA-approved use of injectable IFN-β 1b (Betaseron®, Bayer) (Lin, 1998) to treat MS has been shown to decrease neuronal lesions by 80%, whereas Phase I trials using intramuscular treatment of IFN-α2a (Roferon A, Roche) yields a reduction in the clinical signs of relapse-remitting MS. Moreover, Phase I clinical trials using ingested recombinant human IFN-α2 showed a 30% preservation of b cell function in Type 1 diabetes, as well as a 20% reduction in the signs of symptoms associated with painful and swollen joints in RA (Brod, 2002). Crohn’s disease is another more recently described autoimmune condition which is also thought to arise primarily due to inappropriate chronic T-lymphocyte activation, with tissue damage induced by secondary macrophage activation. The prospect that Crohn’s disease is a form of T-helper type 1 (Th1) cell-dominant autoimmune disease is gaining acceptance, with support from the current use of immunosuppressants, and the possibility of IFN-β utilization as a therapeutics becoming much more probable (Rossi et al., 2009).

IFN-β treatment of MS patients restores immune dysfunction to a degree, but not completely. The incomplete resolution of immune dysfunction by IFNs partly explains their significant, but modest therapeutic effects. This also suggests that there are immune mechanisms in MS and probably other autoimmune diseases that are resistant to IFN therapy. In all autoimmune diseases, abnormalities may exist at several points along the IFN signaling pathway, including molecular defects in the IFN signaling pathways. Currently, many studies are focusing on evaluating ways of differentiating and modulating IFN effects. IFN-β was the first agent to show substantial clinical benefits in the treatment of MS. Many years of experience with these molecules has confirmed clinical efficacy over extended time. In the near future, IFN-based therapeutics will likely continue to play a significant role in treating a variety of autoimmune disorders.

References:

1. Baccala, R., D.H. , and A.N. Theophelopoulos, Interferons as pathogenic effectors in autoimmunity. Immunol Rev, 2005. 204(1):9-26.
2. Brod, S.A., Ingested type I interferon: A potential treatment for autoimmunity. J. Interferon Cytokine Res, 2002. 22:1153-66.
3. Javed, A. and A.T. Reder, Therapeutic role of beta-interferons in multiple sclerosis. Pharmacol Ther, 2005. 110(1):35-56.
4. Lin, L., Betaseron. Dev Biol Stand, 1998. 96:97-104.
5. Pozzilli, C. and L. Prosperini, Clinical markers of therapeutic response to disease modifying drugs. Neurol Sci, 2008. 29(Suppl 2):211-3.
6. Rossi, C.P., et al., IFN-beta1 for the maintenance of remission in patients with Crohn’s disease: results of a phase II dose-finding study. BMC Gastroenterology, 2009. 9(22).
7. Weiner, H.L., A shift from adaptive to innate immunity: a potential mechanism of disease progression in multiple sclerosis. J. Neurol, 2008. 255(Suppl 1):3-11

For more information on the various alleles of IFN alpha 2a, 2b, 4a, 4b, 1 and D, see the references below.

References:

1. Gewart, DR.,  et al. (1995). J. Interferon Cytokine Res. 15(5):403
2. Hussain, M., et al. (1997). J. Interferon Cytokine Res. 17(9):559
3. Hussain, M., et al. (2000). J. Interferon Cytokine Res. 20:763

References:

1. van Pesch et al. (2004). J. Virol. 78(5):8219
2. Hardy et al. (2004). Genomics. 84:331
3. Chen et al. (2004). J. Interferon Cytokine Res. 24:687