Luciferase imaging for evaluation of oncolytic adenovirus replication in vivo


Oncolytic viruses kill cancer cells by tumor-selective replication. Clinical data have established the safety of the approach but also the need of improvements in potency. Efficacy of oncolysis is linked to effective infection of target cells and subsequent productive replication. Other variables include intratumoral barriers, access to target cells, uptake by non-target organs and immune response. Each of these aspects relates to the location and degree of virus replication. Unfortunately, detection of in vivo replication has been difficult, labor intensive and costly and therefore not much studied. We hypothesized that by coinfection of a luciferase expressing E1-deleted virus with an oncolytic virus, both viruses would replicate when present in the same cell. Photon emission due to conversion of D-Luciferin is sensitive and penetrates tissues well. Importantly, killing of animals is not required and each animal can be imaged repeatedly. Two different murine xenograft models were used and intratumoral coinjections of luciferase encoding virus were performed with eight different oncolytic adenoviruses. In both models, we found significant correlation between photon emission and infectious virus production. This suggests that the system can be used for non-invasive quantitation of the amplitude, persistence and dynamics of oncolytic virus replication in vivo, which could be helpful for the development of more effective and safe agents.


Oncolytic virotherapy for treatment of cancer is a rapidly developing field where replicating viruses are utilized as anticancer agents.1 Adenoviruses are useful for cancer gene therapy, because they have demonstrated a good safety profile and lack of oncogenicity in humans, are easy to produce large scale and allow effective gene delivery.2 Moreover, genetic manipulation of adenoviruses is reasonably easy and the capacity for DNA insertion is relatively high. Most importantly, more than 4000 cancer patients have been treated with adenoviruses without known treatment-associated mortality, and efficacy has been confirmed in recent randomized trials.3, 4, 5, 6

Clinical trials have demonstrated that penetration of solid tumor masses is difficult to achieve and an important reason for the limited efficacy seen in many early trials.7 Consequently, replication competent agents such as conditionally replicating adenoviruses (CRAds) have emerged as a strategy to provide local amplification for more effective intratumoral dissemination. Initial clinical trials with pioneering CRAds have demonstrated good safety, but have also highlighted the need for improving efficacy.8 Sufficient transduction of tumor cells continues to be the key variable. Therefore, translational and transductional viral modifications have been employed for increasing specificity and efficacy, and a generation of infectivity-enhanced CRAds is emerging, but has not completed clinical testing yet.6

Nevertheless, several features of solid tumors continue to require further development of agents. For example, intratumoral complexity, including stromal tissues, hypoxic and necrotic areas, hyperbaric and acidic environments present challenges to viral oncolysis.2 These aspects are difficult to test in vitro and therefore means to obtain profound in vivo correlative data are essential. In particular, as most aspects that determine antitumoral efficacy relate to the degree, location and persistence of virus replication, it would be crucial to obtain data on replication kinetics in preclinical models. Further, approaches that would reduce the number of animals required for in vivo work would be ethically and economically advantageous. Therefore, non-invasive imaging of virus replication may be useful for increasing the amount of longitudinal information gained from in vivo experiments.

In this study, we hypothesized that by coinfection of a luciferase expressing E1-deleted virus with an oncolytic virus, both viruses would replicate when present in the same cell. It has been previously shown that E1A proteins produced by a separate virus allows E1-deleted viruses to replicate,9, 10, 11, 12, 13 and as virus DNA replication (including the transgene) is an early event in virus production, the amount of expressed luciferase might correlate with eventual assembly of infectious virions. Therefore, detection of D-Luciferin conversion by luciferase might be useful as a repeatable, non-invasive surrogate for virus replication as a function of time. Here, eight different CRAds as well as control viruses and the replication deficient, luciferase expressing Ad5/3luc1 virus (Table 1) were used in two in vivo experiments with different subcutaneous cancer models.

Table 1 Features and titers of viruses that were used in experiments

Easy, sensitive and non-invasive analysis of in vivo replication would be useful for preclinical evaluation of candidate clinical agents and facilitate overcoming obstacles that currently limit efficacy of oncolytic-viral agents.


Replication kinetics of virus combinations in vitro in HEY ovarian cancer cells

To assess replication kinetics in vitro, HEY cells were infected with a panel of CRAds and control viruses (Table 1) in combination with the replication deficient, luciferase expressing Ad5/3luc1 and analyzed 1, 3, 5 and 7 days later. TCID50 analysis with HEY cell samples revealed an increase of infectious viral particle titers of at least four orders of magnitude from days 1 to 3 for the samples infected with replicating viruses (Figure 1a). TCID50 titers on day 5 and 7 further escalated for all samples infected with replicating viruses, with Ad5/3-Δ24 and Ad5-Δ24pK7 samples having the greatest overall increase from days 1 to 7 of about 7 logs. Samples infected with non-replicating Ad5LacZ and mock samples showed no overall increase although levels were elevated on day 5. Luminescence measurements for luciferase expression showed similar results for Ad5/3-Δ24, Ad5-Δ24pK7 and Ad5-Δ24RGD (Figure 1b). Levels of these samples rose exponentially, so that the overall increase from days 1 to 7 was about three orders of magnitude. Samples infected with Ad5-Δ24E3 showed no overall increase and luciferase expression of Ad300wt and Ad5LacZ samples decreased gradually. Mock-infected cells showed constant background levels. Quantitative polymerase chain reaction (qPCR) was carried out and results were comparable to the findings of TCID50 and luciferase assays (Figure 1c). The samples infected with Ad5/3-Δ24, Ad5-Δ24pK7 and Ad5-Δ24RGD showed over 3 logs increase over 7 days, whereas presence of luciferase gene copies was elevated only slightly for Ad5-Δ24E3 and decreased for Ad300wt and Ad5LacZ.

Figure 1

HEY ovarian cancer cells were infected with indicated viruses in combination with Ad5/3luc1. After 1, 3, 5 and 7 days samples were analyzed by TCID50 assays (a) luciferase expression assays (b) and qPCR for luciferase (c). For all replicating viruses, values of TCID50 assays were plotted against luciferase expression results and statistically analyzed for correlation (d). The same was carried out for TCID50 and qPCR values (e) as well as luciferase expression and qPCR (f).

TCID50 values for all replicating viruses of all four time points were plotted against luciferase light units (RLU) values from luciferase assay and a correlation with R2=0.9908 and P<0.0001 was found (Figure 1d). TCID50 values also showed significant correlation with qPCR values (Figure 1e) with R2=0.9286 and P<0.0001. Furthermore, correlation between RLU and qPCR values was revealed (Figure 1f) with R2=0.9183 and P<0.0001.

Replication kinetics of virus combinations in vitro in 786-O renal cancer cells

786-O cells were infected with various CRAds and control viruses (Table 1) and analyzed after 1, 3, 5 and 7 days. On day 7 cells were mostly dead, therefore it was not able to get reliable result from this day. However, also this cell line showed marked increase of infectious viral particle titers over 5 days for the samples infected with replicating viruses (Figure 2a). Ad5/3-Δ24 and Ad5-Δ24pK7 exhibited once again the greatest overall increase of about 6 logs over 5 days, whereas Ad5LacZ infected samples showed only minimal increase and mock samples stayed approximately on the same levels. Infection with Ad5/3-Δ24 and Ad5-Δ24pK7 also resulted in significantly increased luciferase levels over 5 days whereas Ad5-Δ24RGD and Ad5-Δ24E3 caused no increase (Figure 2b). Samples infected with Ad300wt and Ad5LacZ had decreasing levels and mock-infected cells showed only baseline light emission. qPCR with luciferase specific primers showed that the number of luciferase gene copies rose in the cells infected with Ad5/3-Δ24, Ad5-Δ24pK7 and Ad5-Δ24RGD by approximately three orders of magnitude over 5 days. Samples infected with Ad5-Δ24E3 exhibited an increase of about 1 log whereas infection with Ad300wt and Ad5LacZ showed no increase.

Figure 2

786-O renal cancer cells were infected with indicated viruses in combination with Ad5/3luc1. After 1, 3 and 5 days samples were analyzed by TCID50 assays (a) luciferase expression assays (b) and qPCR for luciferase (c). For all replicating viruses, values of TCID50 assays were plotted against luciferase expression results and statistically analyzed for correlation (d). The same was carried out for TCID50 and qPCR values (e) as well as luciferase expression and qPCR (f).

TCID50 values for all replicating viruses of all three time points were plotted against RLU values from luciferase assay and a correlation with R2=0.9794 and P<0.0001 was found (Figure 2d). Correlation was also found for TCID50 and qPCR results (Figure 2e) with R2=0.9915 and P<0.0001. Moreover, luciferase assay values RLU and qPCR results correlated with R2=0.9868 and P<0.0001.

Replication kinetics of single-component luciferase expressing replicating adenovirus in vitro in HEY and 786-O cells

To compare the two-component system to a single virus, the experiments were repeated with Ad5luc3, a replicating adenovirus that has E3 replaced with an SV-40 promoter – luciferase expression cassette. With HEY cells, viral particle counts, luciferase activity and quantitative adenovirus genomes all increased two-threefold between days 1–3 (Figure 3a–c). Similar data were obtained in 786-O cells (Figure 3d–f). In general, Ad5luc3 seemed to replicate slightly slower than the CRAds with intact E3 used in the combination experiments (Figures 1 and 2). This is probably due to slower replication of E3-deleted adenoviruses (the other CRAds were E3 positive) and also the activity of a heterologous promoter working against virus replication, as reported previously.14, 15, 16 Most importantly, the single-component system confirmed the validity of the more widely applicable coinfection strategy, in that the three measured variables correlated with each other.

Figure 3

HEY (ac) and 786-O cells (df) were infected with replicating Ad5luc3 or replication deficient Ad5luc1. After 1, 3, 5 and 7 days, samples were analyzed by TCID50 assays (a, d) luciferase expression assays (b, e) and qPCR for luciferase (c, f).

Bioluminescence imaging of coinfected ovarian tumors

To analyze factors associated with treatment resistance, a subcutaneous model featuring Hey ovarian adenocarcinoma cells was utilized in the first in vivo experiment. These cells grow rapidly and have been difficult to treat in previous adenoviral gene therapy experiments.17, 18 CRAds featuring Cox-2 promoters and/or E1A region mutations and control viruses (Table 1) were coinjected with replication deficient Ad5/3luc1 into subcutaneous tumors. Three days after first injection of viruses, all treated tumors showed photon emission, suggesting successful transduction of tumor cells and subsequent in vivo conversion of D-Luciferin by luciferase (Figure 4). When imaging was redone on day 9, many tumors continued to display photon emission, suggesting presence of viruses in the tumors. In some tumors, photon emission had increased, suggesting persistent CRAd replication, whereas in other tumors photon emission decreased or ceased. To evaluate if this was due to death of tumor cells or attenuation of replication, another injection of viruses was carried out on day 10. Larger tumors usually again effectively produced luciferase, whereas smaller tumors continued to exhibit low photon emission. This suggested that the smaller tumors had few remaining viable tumor cells, whereas the larger tumors could again sustain virus replication.

Figure 4

Photon emission from mice with subcutaneous Hey ovarian adenocarcinoma tumors. When tumors reached a volume of ca. 65 mm3, they were injected with the indicated virus+luciferase coding Ad5/3luc1 (days 0, 1, 2 and 10). Mock mice received no virus. Photon counts were obtained on days 3, 9 and 13 as shown. Half of the tumors were collected for titer determination on day 3 and the rest on day 17.

Therapeutic efficacy, bioluminescence imaging and analysis of infectious virus production in ovarian cancer model

Due to the refractory nature of Hey tumors, many tumors grew bigger throughout the experiment. However, some viruses, such as Ad5/3-Δ24 and Ad5/3Cox2Ld24, were able to slow down tumor growth (Figure 5a) and some tumors even disappeared by the end of the experiment. Antitumor efficacy was significant for Ad5/3-Δ24 and Ad5/3Cox2Ld24 (P=0.0024 and 0.0131) when compared to mock group.

Figure 5

Tumor volume (a), photon emission (b) and infectious virus production (c). Hey ovarian adenocarcinoma tumors were injected with the indicated virus+Ad5/3luc1 (days 0, 1, 2 and 10). Mock mice received no virus. Tumor volume at day 0 was set 100%. Error bars indicate s.e.m.. *P<0.05, **P<0.01 when compared to Ad5LacZ. (df) Correlation between tumor size, photon emission and infectious virus production in the Hey tumor experiment. Relative volume of CRAd injected tumors was plotted against infectious particles (TCID 50) and a trend line was drawn (d). The same was carried out for photons vs TCID50 (e) and relative tumor volume vs total flux (f).

Photon emission values on day 3 were quite similar between the CRAd coinfected groups, whereas values for replication deficient Ad5LacZ coinfected tumors were low (Figure 5b). This suggests that presence of a replicating virus with an E1-deleted virus allows replication of both viruses. On day 9, light emission was less pronounced in many of the CRAd treated groups, especially when treated with Ad5/3-Δ24 and Ad5/3Cox2Ld24. On day 13 (3 days after the last virus injection) all groups treated with CRAds+Ad5/3luc1 displayed higher light emission than 4 days before.

The amounts of infectious viruses produced in the tumors were analyzed by TCID50 assays (Figure 5c). Titers were very low for Ad5LacZ+Ad5/3luc1 suggesting lack of replication in the absence of E1A. Values for Ad5/3-Δ24 were four times higher than for the other CRAds with the tumors that were harvested on day 3. Low TCID50 values were seen for Ad5/3Cox2LE1 and Ad5/3Cox2Ld2d24 on day 17, whereas values for Ad300wt and also for Ad5/3-Δ24 were relatively high. Groups treated with CRAds that contain the Cox2 promoter showed lower TCID50 values than tumors treated with other replicating viruses (Ad5/3-Δ24 and Ad300wt).

Correlation of tumor volume, infectious virus production and photon emission in ovarian cancer model

To determine whether there is a relation between tumor size and virus replication (TCID50), values were displayed against each other in a scatter plot (Figure 5d). Statistical analysis revealed a significant correlation (P=0.018). When photon emission and TCID50 values were plotted against each other (Figure 5e) a significant correlation could be seen as well (P=0.029). Relative tumor volume also seemed to correlate with photon emission (Figure 5f, P=0.050).

Bioluminescence imaging of coinfected subcutaneous renal tumors

To investigate the generalizability of the findings, a tumor model with 786-O renal cancer cells was chosen and CRAds with different capsid modifications and control viruses (Table 1) were coinjected with Ad5/3luc1. All virus injected tumors showed relatively high photon values on day 3, whereas on day 9 values were lower (Figure 6) suggesting that most of the tumor tissue was eliminated by this time.

Figure 6

Photon emission from mice with subcutaneous 786-O renal cancer tumors. When tumors reached a volume of ca. 65 mm3, they were injected with the indicated virus+Ad5/3luc1 (days 0, 1 and 2). Mock mice received no virus. Photon counts were obtained on days 3 and 9 as shown. Half of the tumors were collected for titer determination on day 3 and the rest on day 9.

Therapeutic efficacy, bioluminescence imaging and analysis of infectious virus production in renal cancer model

786-O tumors responded well to virus treatment as all tumors injected with replication competent viruses shrank, whereas mock-injected tumors grew (Figure 7a). All the CRAds used in this experiment had a significant antitumor effect compared to Ad5LacZ (P-values on day 9: Ad5/3-Δ24=0.002, Ad5-Δ24pK7=0.004, Ad5-Δ24RGD=0.007 and Ad5-Δ24E3=0.01).

Figure 7

Tumor volume (a), photon emission (b), and infectious virus production (c) in the renal cancer model. 786-O renal cell tumors were injected with the indicated virus + Ad5/3luc1 (days 0, 1 and 2). Mock mice received no virus. Different scales are provided for days 3 and 9 (b, c) to allow meaningful comparison of groups. Error bars indicate standard error of the mean. *P<0.05, **P<0.01 when compared to Ad5LacZ. (df) Correlation between tumor size, photon emission and infectious virus production in the 786-O renal cell cancer experiment. Relative volumes of CRAd injected tumors were plotted against infectious particles (TCID 50) and a trend line was drawn (d). The same was carried out for photons vs TCID50 (e) and relative tumor volume vs total flux (f).

Photon emission values on day 3 were highest for the Ad5/3-Δ24+Ad5/3luc1 group and lowest for the replication deficient Ad5LacZ+Ad5/3luc1 which indicates successful coreplication of Ad5/3luc1 with the CRAds (Figure 7b). On day 9 photon emissions had decreased at least 100-fold in all virus injected groups, due to killing of tumor cells. Nevertheless, the Ad5/3-Δ24 group continued to display the highest values.

TCID50 analysis of tumors harvested on day 3 showed high values for Ad300wt and Ad5/3-Δ24, whereas tumors injected with replication deficient Ad5LacZ exhibited low values (Figure 7c). Results for the samples collected on day 9 were markedly decreased but still tumors transduced with Ad300wt and Ad5/3-Δ24 had the highest values.

Correlation of tumor volume, infectious virus production and photon emission in the renal cell tumor model

To assess a relation between TCID50 and tumor volume, values were plotted against each other (Figure 7d) and a significant correlation was found (P=0.0011). The same was carried out for TCID50 values and total flux (Figure 7e) and a significant correlation could be seen as well (P=0.0003). Also parameters for total flux and tumor volume correlated with each other (P=0.0024, Figure 7f).

Correlation of virus replication and antitumor effect at different time points

To investigate correlation of early virus replication to subsequent antitumor activity, flux on day 3 was plotted against tumor size on day 9. However, low correlation was seen (not significant, not shown) for both renal cancer and ovarian cancer model.

Comparison of a single-component luciferase expressing replicating adenovirus to the dual infection strategy in vivo

786-O tumors were injected with Ad5luc3 alone or Ad300wt in combination with Ad5/3luc1 and similar amounts of light were emitted on days 3 and 9 (Figure 8a). Total flux values were calculated and plotted as a function of time (Figure 8b). Furthermore, production of new infectious virions was similar although the coinfection yielded slightly more virus, probably due to higher replicativity of Ad300wt in comparison to E3-deleted Ad5luc3 (Figure 8c).

Figure 8

Mice with subcutaneous 786-O renal tumors were infected with Ad5luc3 or Ad300wt+Ad5/3luc1. Three or 9 days later, mice were imaged for luciferase (a) and total photon emission was determined (b). Half of the mice were killed on days 3 and 9 and tumors were analyzed for functional VPs (c). Bars indicate standard error.


Adenoviruses are promising tools for cancer therapy, and their safety and efficacy has been shown in recent clinical trials.2, 3, 4, 5, 6 However, clinical data have also revealed that eradication of advanced tumors is difficult with non-replicating viruses, and the best results have been obtained in combination treatment regimens,4, 5 or with minimal residual disease.3 Therefore, replication competent oncolytic viruses were developed and the first generation has now been evaluated in clinical trials. Although safety has been excellent and some responses have been seen, single-agent efficacy has been less than impressive.8 Reasons for the discrepancy between preclinical and clinical data probably include poor transduction of sufficient numbers of tumor cells and unsatisfactory intratumoral dissemination due to intratumoral barriers. Advanced clinical tumors are larger, slower growing and more complex than preclinical models and the viruses tested heretofore are not very potent. Consequently, a new generation of infectivity enhanced, highly potent CRAds is emerging.19, 20 Although clinical data on these agents is still pending, it is increasingly recognized that the intratumoral environment of advanced solid tumors is quite complex, and improved means for in vivo analysis are required. In particular, currently available methods rarely allow longitudinal correlation of virus replication to antitumor responses.

To address this problem, we established a system to monitor viral replication with a non-invasive technique, which allows multiple measurements with the same test animals. A replication-deficient, luciferase expressing virus was coinjected with different CRAds to investigate whether light emission values could be used as a surrogate for killing of test animals for analysis of virions produced.

In vitro, good correlation of the measured parameters was found (Figures 1d–f and 2d–f). TCID50 titers increased by 6–7 log (Figures 1a and 2a), whereas luciferase copy number generally increased by ca. 4 log (Figures 1c and 2c), which suggests that coreplication of the viruses was not complete. Alternatively, methodological issue may have impacted this quantitative difference: TCID50 measures total functional particles, whereas qPCR determines total genomes of the non-replicating virus. However, in a one-component system we saw quite similar results (Figure 3), suggesting that luciferase imaging could track replication regardless of E1 provided in trans or by the same virus.

The in vivo experiments showed good correlation between production of infectious virions and photon emission values (Figures 5e, 7e and 8). This suggests that in vivo luciferase imaging can be used to estimate virus replication in live animals. Interestingly, we also saw correlation between tumor volume and infectious virus production, and between tumor volume and photon emission. These findings initially seemed partly counterintuitive, as more replication would be expected to result in more antitumor activity and therefore smaller tumors. Although daily analysis of virus production and photon emission was not possible, we assume that virus replication was highest at early time points (days 0–2) in the tumors where the best responses were seen (found along the y-axis in Figures 5d and f, 7d and f). Therefore, correlation of tumor size with photon and virus counts occurs mostly at later time points and presumably reflects a situation where the tumor is not completely eradicated (due to intratumoral barriers), but antitumor activity mediated by virus replication persists (Figure 5d and f). In the 786-O renal cancer model, it is noteworthy that the trend line crosses the y-axis not at the origo but higher up (Figure 7d and f). This suggests that in the smallest tumors, little if any virus replication existed, because no viable tumor cells were present anymore.

An advantage of this system is that construction of marker gene expressing variants of the oncolytic viruses of interest is not necessary, as a luciferase expressing virus can be coinfected with any oncolytic virus. For example, oncolytic viruses with various configurations of promoters and E1-deletions could be tested and compared for transcriptional activity and the degree of subsequent virus replication obtained (ovarian cancer model, Figures 4 and 5). In our experiments, Ad5/3-Δ24 and Ad5/3Cox2Ld24 seemed the most useful viruses. By utilizing a cell line that theoretically should not support replication of a CRAd, it might be possible to obtain information about selectivity. For example, if Cox-2 promoter containing CRAds are studied, a Cox-2-negative cell line could be used.

Alternatively, different viral capsid constructs can be compared with this system (renal cancer model, Figures 6 and 7). Provided the compared viruses are otherwise isogenic, results would then allow to draw conclusions about the transductional performances of the capsid modified viruses. In our experiments with a subcutaneous renal cell cancer model, all three capsid modified CRAds (Ad5-Δ24RGD, Ad5-Δ24pK7, Ad5/3-Δ24) were effective.

In the optimum situation, the capsid of the luciferase-expressing virus would be identical with the replication competent virus (Figures 4 and 5). Otherwise, there is a theoretical possibility that non-identical populations of tumor cells are transduced with the luciferase expressing virus and the replication competent one. However, our data suggest that this is not a concern, as logical results were obtained in Figures 6 and 7. In general, high TCID50 correlated with high photon emission and both seemed to correlate with antitumor response. Nevertheless, on day 3, Ad5/3-Δ24 displayed the highest photon emission (Figure 7b), but Ad300wt displayed higher TCID50 values (Figure 7c). Because the coinfected virus was Ad5/3luc1 and therefore had the same capsid as Ad5/3-Δ24, it may have led to a slight advantage with regard to photon emission vs TCID50 for this CRAd.

In summary, we have shown here that the proposed system of coinjecting a luciferase expressing, replication deficient adenovirus with a replication competent adenovirus allows determination of replication non-invasively. Moreover, the analysis can be repeated in the same animals to obtain longitudinal data. Thus, this system might be useful for replacing or complementing more laborious TCID50 assays with tissue samples. Because killing of animals is required for TCID50 analysis, this system would reduce the number of required test animals, which would be ethically and economically beneficial. Although other imaging approaches may be useful to obtain data complementary to this approach,21, 22, 23 there is unfortunately a paucity of methods with wide applicability without the need for constructing novel constructs. Because an imaging moiety cannot be easily genetically included in all oncolytic agents due to genome size, packaging, safety and efficacy issues, this coinfection strategy might be frequently useful as it is theoretically combinable with any oncolytic adenovirus. Eventually, this system could improve the analytical methods available for preclinical evaluation of oncolytic viruses, which in turn could lead to development of more effective treatment agents for patients currently faced with disease refractory to available modalities.

Materials and methods


The main features of the viruses used in this study can be found in Table 1. All viruses contain an intact E3 region, except Ad5luc3 which has E3 replaced with an SV-40 – luciferase expression cassette.24 Viruses were produced according to standard methodology as referenced, purified with double CsCl centrifugation, and titered with OD260 and plaque assay.

Ad300wt was obtained from American type culture collection (ATCC, Manassas, VA, USA).

Cell lines

HEY ovarian cancer cells25 were obtained from Dr Wolf (MD Anderson Cancer Center, Houston, TX, USA), 786-O renal cancer cells and 293 cells were purchased from ATCC (CRL-1932 and CRL-1573). All cell lines were cultured under recommended conditions.

HEY cells have been described earlier to be Cox2-positive.26

In vitro assays

A total of 250 000 HEY and 786-O cells per well were infected with 1.25 × 106 virus particles (VP) of the respective CRAd (Table 1) or control virus in combination with 1.25 × 106 VPs of Ad5/3luc1. For the assay with the replicating, luciferase-expressing virus, 1.25 × 106 VPs of either Ad5luc3 or Ad5luc1 alone was used to infect cells. On days 1, 3, 5 and 7 after infection cells were harvested and analyzed by TCID50 assays with 293 cells (Adeasy application manual, Qbiogene, Carlsbad, CA, USA), luciferase expression (Luciferase Assay System, Promega, Madison, WI, USA) and qPCR for luciferase gene. For TCID50 assays cell samples were freeze-thawed thrice to set free virus, for luciferase expression assays the provided reporter lysis buffer was used and for qPCR DNA was extracted from samples using QIAamp DNA Mini Kit (Qiagen, Valencia, CA, USA). Analyses were carried out in triplicates for each virus, day and analytical method.

qPCR for luciferase

Luciferase specific primers (5′-IndexTermGAAATCCCTGGTAATCCGTT-3′ and 5′-IndexTermATCACAGAATCGTCGTATGC-3′) were used to perform real-time quantitative PCR with a SYBR green assay using a RotorGene system. The efficiency of the reaction was assumed to be 2, therefore values were calculated with the formula: 2(MockCycleNumber−SampleCycleNumber).

Animal models

All experiments were approved by Experimental Committee of the University of Helsinki and the Provincial Government of Southern Finland. About 4–5-weeks old female nude mice were purchased from Taconic (Ejby, Denmark). For the ovarian cancer model, mice were subcutaneously injected with 5 × 106 Hey ovarian adenocarcinoma cells in both flanks. When tumors reached a volume of ca. 65 mm3, mice were randomized into seven groups receiving indicated virus or no virus intratumorally on three consecutive days (days 0, 1 and 2) and again on day 10 (n=8 tumors per group). Each injection contained 3 × 108 VPs of replicating virus and 3 × 108 VP of Ad5/3luc1. Half of the mice were killed on day 3 and the rest on day 17. All tumors were collected and stored at −80°C.

For the renal cancer model, subcutaneous tumors were induced by injecting 5 × 106 786-O renal cancer cells as above. Mice were randomized into six groups and treated on three consecutive days (days 0, 1 and 2) with 3 × 108 VPs of replicating virus and 3 × 108 VPs of Ad5/3luc1 (n=8 tumors per group). For correlation of single to dual component in vivo imaging, the coinfection strategy was compared to 3 × 108 VPs of Ad5luc3 injected intratumorally on three consecutive days. Half of the mice were killed on day 3 and the rest on day 9.

Tumor size was followed and volumes were calculated using the formula: (larger diameter) × (smaller diameter)2 × 0.52. All tumors were collected and stored at −80°C.


On days 3 and 9 (in all experiments) and 13 (only in the ovarian cancer model) after the first virus injection, 4.5 mg of D-Luciferin (Promega, Madison, WI, USA) in 100 μl 0% Roswell Park Memorial Institute growth medium was injected intraperitoneally. After 10 min, images were captured with the IVIS imaging system series 100 using Living Image v2.5 software (Xenogen, Alameda, CA, USA) and photon emission values were calculated as recommended by the manufacturer.

Quantitation of infectious particles of in vivo samples

Tumors were minced, suspended in 1 ml 0% Dulbecco's modified Eagle's medium growth medium and freeze–thawed three times.27 The supernatant was then used for duplicate 10 day TCID50 assays with 293 cells (Adeasy Application Manual, QBiogene).

Statistical analysis

Significance of antitumor efficacy for the ovarian cancer model experiment was calculated using a non-parametric change-point test to show a systematic change in the pattern of observations as opposed to fluctuation due to chance. The Proc Mixed procedure in 18SAS v.6.12 (SAS Institute, Cary, NC, USA) was used to examine the effects of group and time on tumor growth. Pairwise comparisons were performed to compare groups.

RLU, qPCR, TCID50, in vivo photon emission and tumor volume values for all CRAds were entered into scatter plots and correlation (two-tailed Pearson's t-test) was determined using GraphPad Prism 4 software (GraphPad Software Inc., San Diego, CA, USA).


  1. 1

    Kirn D, Martuza RL, Zwiebel J . Replication-selective virotherapy for cancer: biological principles, risk management and future directions. Nat Med 2001; 7: 781–787.

  2. 2

    Kanerva A, Hemminki A . Adenoviruses for treatment of cancer. Ann Med 2005; 37: 33–43.

  3. 3

    Immonen A, Vapalahti M, Tyynela K, Hurskainen H, Sandmair A, Vanninen R et al. AdvHSV-tk gene therapy with intravenous ganciclovir improves survival in human malignant glioma: a randomised, controlled study. Mol Ther 2004; 10: 967–972.

  4. 4

    Xia ZJ, Chang JH, Zhang L, Jiang WQ, Guan ZZ, Liu JW et al. Phase III randomized clinical trial of intratumoral injection of E1B gene-deleted adenovirus (H101) combined with cisplatin-based chemotherapy in treating squamous cell cancer of head and neck or esophagus. Ai Zheng 2004; 23: 1666–1670.

  5. 5

    Peng Z . Current status of gendicine in China: recombinant human Ad-p53 agent for treatment of cancers. Hum Gene Ther 2005; 16: 1016–1027.

  6. 6

    Kanerva A, Hemminki A . Modified adenoviruses for cancer gene therapy. Int J Cancer 2004; 110: 475–480.

  7. 7

    Puumalainen AM, Vapalahti M, Agrawal RS, Kossila M, Laukkanen J, Lehtolainen P et al. Beta-galactosidase gene transfer to human malignant glioma in vivo using replication-deficient retroviruses and adenoviruses. Hum Gene Ther 1998; 9: 1769–1774.

  8. 8

    Kirn D . Clinical research results with dl1520 (Onyx-015), a replication-selective adenovirus for the treatment of cancer: what have we learned? Gene Therapy 2001; 8: 89–98.

  9. 9

    Hemminki A, Dmitriev I, Liu B, Desmond RA, Alemany R, Curiel DT . Targeting oncolytic adenoviral agents to the epidermal growth factor pathway with a secretory fusion molecule. Cancer Res 2001; 61: 6377–6381.

  10. 10

    Alemany R, Lai S, Lou YC, Jan HY, Fang X, Zhang WW . Complementary adenoviral vectors for oncolysis. Cancer Gene Ther 1999; 6: 21–25.

  11. 11

    Umeoka T, Kawashima T, Kagawa S, Teraishi F, Taki M, Nishizaki M et al. Visualization of intrathoracically disseminated solid tumors in mice with optical imaging by telomerase-specific amplification of a transferred green fluorescent protein gene. Cancer Res 2004; 64: 6259–6265.

  12. 12

    Thorne SH, Tam BY, Kirn DH, Contag CH, Kuo CJ . Selective intratumoral amplification of an antiangiogenic vector by an oncolytic virus produces enhanced antivascular and anti-tumor efficacy. Mol Ther 2006; 13: 938–946.

  13. 13

    Habib NA, Mitry R, Seth P, Kuppuswamy M, Doronin K, Toth K et al. Adenovirus replication-competent vectors (KD1, KD3) complement the cytotoxicity and transgene expression from replication-defective vectors (Ad-GFP, Ad-Luc). Cancer Gene Ther 2002; 9: 651–654.

  14. 14

    Suzuki K, Alemany R, Yamamoto M, Curiel DT . The presence of the adenovirus E3 region improves the oncolytic potency of conditionally replicative adenoviruses. Clin Cancer Res 2002; 8: 3348–3359.

  15. 15

    Hemminki A, Wang M, Hakkarainen T, Desmond RA, Wahlfors J, Curiel DT . Production of an EGFR targeting molecule from a conditionally replicating adenovirus impairs its oncolytic potential. Cancer Gene Ther 2003; 10: 583–588.

  16. 16

    Hakkarainen T, Hemminki A, Curiel DT, Wahlfors J . A conditionally replicative adenovirus that codes for a TK-GFP fusion protein (Ad5Delta24TK-GFP) for evaluation of the potency of oncolytic virotherapy combined with molecular chemotherapy. Int J Mol Med 2006; 18: 751–759.

  17. 17

    Hemminki A, Zinn KR, Liu B, Chaudhuri TR, Desmond RA, Rogers BE et al. In vivo molecular chemotherapy and noninvasive imaging with an infectivity-enhanced adenovirus. J Natl Cancer Inst 2002; 94: 741–749.

  18. 18

    Kanerva A, Wang M, Bauerschmitz GJ, Lam JT, Desmond RA, Bhoola SM et al. Gene transfer to ovarian cancer versus normal tissues with fiber-modified adenoviruses. Mol Ther 2002; 5: 695–704.

  19. 19

    Bauerschmitz GJ, Lam JT, Kanerva A, Suzuki K, Nettelbeck DM, Dmitriev I et al. Treatment of ovarian cancer with a tropism modified oncolytic adenovirus. Cancer Res 2002; 62: 1266–1270.

  20. 20

    Kanerva A, Zinn KR, Chaudhuri TR, Lam JT, Suzuki K, Uil TG et al. Enhanced therapeutic efficacy for ovarian cancer with a serotype 3 receptor-targeted oncolytic adenovirus. Mol Ther 2003; 8: 449–458.

  21. 21

    Ono HA, Le LP, Davydova JG, Gavrikova T, Yamamoto M . Noninvasive visualization of adenovirus replication with a fluorescent reporter in the E3 region. Cancer Res 2005; 65: 10154–10158.

  22. 22

    Richard JC, Factor P, Ferkol T, Ponde DE, Zhou Z, Schuster DP . Repetitive imaging of reporter gene expression in the lung. Mol Imaging 2003; 2: 342–349.

  23. 23

    Nanda D, de Jong M, Vogels R, Havenga M, Driesse M, Bakker W et al. Imaging expression of adenoviral HSV1-tk suicide gene transfer using the nucleoside analogue FIRU. Eur J Nucl Med Mol Imaging 2002; 29: 939–947.

  24. 24

    Krasnykh VN, Mikheeva GV, Douglas JT, Curiel DT . Generation of recombinant adenovirus vectors with modified fibers for altering viral tropism. J Virol 1996; 70: 6839–6846.

  25. 25

    Buick RN, Pullano R, Trent JM . Comparative properties of five human ovarian adenocarcinoma cell lines. Cancer Res 1985; 45: 3668–3676.

  26. 26

    Bauerschmitz GJ, Guse K, Kanerva A, Menzel A, Herrmann I, Desmond RA et al. Triple-targeted oncolytic adenoviruses featuring the cox2 promoter, E1A transcomplementation, and serotype chimerism for enhanced selectivity for ovarian cancer cells. Mol Ther 2006; 14: 164–174.

  27. 27

    Kanerva A, Zinn KR, Peng KW, Ranki T, Kangasniemi L, Chaudhuri TR et al. Noninvasive dual modality in vivo monitoring of the persistence and potency of a tumor targeted conditionally replicating adenovirus. Gene Therapy 2005; 12: 87–94.

  28. 28

    Kanerva A, Mikheeva GV, Krasnykh V, Coolidge CJ, Lam JT, Mahasreshti PJ et al. Targeting adenovirus to the serotype 3 receptor increases gene transfer efficiency to ovarian cancer cells. Clin Cancer Res 2002; 8: 275–280.

  29. 29

    Yotnda P, Zompeta C, Heslop HE, Andreeff M, Brenner MK, Marini F . Comparison of the efficiency of transduction of leukemic cells by fiber-modified adenoviruses. Hum Gene Therapy 2004; 15: 1229–1242.

  30. 30

    Ranki T, Kanerva A, Ristimaki A, Hakkarainen T, Sarkioja M, Kangasniemi L et al. A heparan sulfate-targeted conditionally replicative adenovirus, Ad5.pk7-Delta24, for the treatment of advanced breast cancer. Gene Therapy 2007; 14: 58–67.

  31. 31

    Suzuki K, Fueyo J, Krasnykh V, Reynolds PN, Curiel DT, Alemany R . A conditionally replicative adenovirus with enhanced infectivity shows improved oncolytic potency. Clin Cancer Res 2001; 7: 120–126.

Download references


This study was supported by Helsinki Graduate School in Biotechnology and Molecular Biology, Emil Aaltonen Foundation, EU FP6 THERADPOX and APOTHERAPY, HUCH Research Funds (EVO), Sigrid Juselius Foundation, Academy of Finland, Finnish Cancer Society, University of Helsinki, Sohlberg Foundation, Instrumentarium Research Fund, the Finnish Oncology Association and Deutsche Forschungsgemeinschaft (DFG).

Author information

Correspondence to A Hemminki.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Guse, K., Dias, J., Bauerschmitz, G. et al. Luciferase imaging for evaluation of oncolytic adenovirus replication in vivo. Gene Ther 14, 902–911 (2007).

Download citation


  • non-invasive imaging
  • adenovirus
  • luciferase
  • cancer
  • replication
  • CRAd

Further reading