Drugs

Eight methylxanthine drugs were used in the study: aminophylline, caffeine, diprophylline, etophylline, IBMX, theophylline from Sigma and doxofylline and pentoxifylline from Abcam. Metoprolol, a β-blocker used for the cardiovascular study was obtained from Sigma.

Zebrafish maintenance and collection of eggs

Zebrafish were maintained and used in accordance with the Italian and European rules on animal use following protocols approved by the local Committee (OPBA protocol nr 211B5.24) and authorized by the Ministry of Health (authorization number 393/2017-PR).

Adult AB wild type and transgenic line tg (Bmp:EGFP) [33] zebrafish were raised and maintained in 14: 10 h light: dark cycle at 28 °C in a circulating system maintained at pH 7.0 and conductivity between 400 and 500 μs. Fishes were fed three time per day with a combination of granular food (from special diet services, SDS, Witham, UK) in the morning and evening and artemia freshly prepared in the laboratory in the afternoon (cysts bought from SDS, Witham, UK).

Adult zebrafish were used for egg production. Male and female were put in the breeding tank overnight and next morning fresh spawned embryos were collected. Any unfertilized eggs, dead or bad embryos were removed and the good quality embryos at 1-2 cell stage were selected for microinjection.

Microinjection

Evaluation of mortality and general morphology

Evaluation of cardiovascular effects

After performing microinjection as described above, structural cardiac defect and functional cardiac effects of the methylxanthine drugs were evaluated together with other cardiovascular endpoints.

Structural cardiac defect

The structural cardiac abnormality was identified under direct visualization of the heart of zebrafish embryos using Leica MZ 16F Stereomicroscope (Germany) at 80X magnification.

Evaluation of cardiovascular endpoints

In addition to the structural cardiac defects, we also evaluated the presence of several cardiovascular endpoints which included pericardial edema, decreased or absent blood flow, arrhythmia, hemorrhage and thrombosis. The total number of embryos developing each of these endpoints with all the eight methylxanthines was documented and the resulting percentage calculated for each of the endpoints at 48 hpf.

To see the fate of the embryos with cardiac defects beyond 72 hpf, one selected concentration of each of the eight compounds was injected into the embryos at 1 cell stage. The embryos with cardiac defects were then evaluated for mortality till 5 dpf.

Functional cardiac effect

To study the functional cardiac effects of methylxanthines, the heart rate of normal looking embryos was measured at 48 hpf. Normal looking embryos implied the embryos with intact structural integrity including the normal cardiac strucuture and cardiovascular endpoints. For each drug, ten embryos were randomly selected from a pool of normal looking embryos and the heart beat of each of the embryos was counted manually for 30 s under Leica MZ16F stereomicroscope (Germany) in quiet condition and at optimal temperature. The heart beat per 30 s was multiplied by 2 to obtain the heart beat per minute i.e. heart rate. Metoprolol, a beta blocker which decreases the heart rate in human, was used as an internal control and embryos injected with sterile water was used as negative control.

To further understand if the functional cardiac effect was reversible or irreversible, we performed the same experiment in the Tg(Bmp:EGFP) transgenic line with the embryos injected with one concentration of each drug. BMP transgenic line allows for the easy visualization of heartbeat, especially at 72 and 120 hpf when the pigmentation might make it difficult to count the heart beat in the wild type embryos. The heart rate was measured starting from 48 hpf till 120 hpf, at 24-h interval. The procedure for the measurement of heart beat was same as described earlier. However, for the measurement of heart beat at 72, 96 and 120 hpf, embryos were stabilized by briefly placing them for 10-15 s in low dose tricaine (0.16 mg/mL) before measuring the heart beat.

Statistical analysis

All the experiments were repeated at least three times and the data are reported in Mean ± S.D. One way ANOVA with Dunnett’s test was used for the comparison of the heart rate of the treated embryos with control. Statistical analysis was done in graph pad prism 5 (La Jolla, CA USA) with p value <0.05 considered statistically significant.

Embryotoxicity

As a first step in the characterization of the toxicity of methylxanthine compounds, we studied the lethality and phenotypic effects.

Five increasing doses/graded doses of eight methylxanthines: aminophylline, caffeine, diprophylline, doxofylline, etophylline, 3-isobutyl-1-methylxanthine (IBMX), pentoxifylline and theophylline; were injected into the yolk of 1-2 cell zebrafish embryos (Table 1) and the survival and phenotype analyzed at specific timepoints. The doses were selected in such a way that they fall within per oral therapeutic range in human. As the weight of zebrafish embryos is roughly 1.2 mg, administering 1 ng of a drug in one embryo (1 ng/1.2 mg) is approximately equal to 1 mg/kg of drug. Therefore, the doses of drugs used in microinjection were in the range of 1-20 mg/kg and most of these drugs have per oral therapeutic dose that falls within this range in humans.

Aminophylline, caffeine, doxofylline, IBMX, pentoxifylline and theophylline treated embryos showed dose dependent survival at 72 hpf, whereas the survival of diprophylline and etophylline treated embryos was independent of the doses (Fig. 1). IBMX was the most toxic/active compound: a large percentage of the embryos (58%) showed mortality when injected with 1 ng of the drug. Diprophylline was the least active methylxanthine; with more than 90% of embryos surviving at dose as high as 20 ng/embryo. Pentoxifylline and caffeine also produced >50% mortality, although they required higher doses than IBMX, i.e. 3 ng and 2.5 ng per embryo respectively to produce the equivalent mortality.

To simplify the comparison, we observed the dose at which there was a minimum of 20% mortality for each drug. Methylxanthines like IBMX, caffeine, pentoxifylline and theophylline were highly active with <1 ng needed to induce 20% mortaility. Aminophylline, etophylline and doxofylline were moderately active with 2.5-5 ng required to induce 20% mortality and diprophylline was a non toxic compound which didn’t induce 20% mortality even with the highest dose of 20 ng. To further compare the toxicity of each individual compound, we calculated the lethal dose at which there was 50% mortality (LD₅₀) for each drug using linear interpolation and extrapolation. Our results showed that IBMX was the most toxic drug (LD50 = 0.91 ng) followed by caffeine (2.4 ng), pentoxifylline (2.76 ng), theophylline (3.4 ng) and aminophylline (6.15 ng), etophylline (9.8 ng),doxofylline (29.3 ng) and diprophylline (167.5 ng).

General Teratogenicity and general morphological defects

After evaluating the mortality, the alive embryos were observed for morphological defects at 48 and 72 hpf. The embryos with morphological defects were identified and classified into mild and severe phenotype based on the presence of different morphological endpoints (Table 2). In particular, embryos with defective heart formation, irregular heartbeat, decreased blood circulation, mild pericardial and yolk sac edema were categorized as mild phenotype. Embryos with perturbed anterior-posterior axis development with reduced tail detachment and poor formation of somites, defective heart formation, irregular heartbeat, absent blood circulation, severe pericardial and yolk sac edema were categorized as severe phenotype.

The percentage of embryos that showed morphological defects with the highest concentration of drug varied with each compound. While the lowest dose used produced morphological defects in <10% of the embryos, the highest concentration produced morphological defects in 50% of embryos with 2.5 ng of caffeine, 44% with 1 ng of IBMX, 38% with 3 ng of pentoxifylline, 29% with 5 ng of aminophylline, 24% with 1 ng of theophylline, 23% with 7.5 ng of doxofylline, 15% with 5 ng of etophylline and 5% with 20 ng of diprophylline. The embryos injected with the positive control DCA showed 47% morphological defects, while less than 1% of negative controls injected with sterile water showed morphological defects (Fig. 2). These embryos predominantly had pericardial edema, abnormal blood circulation, yolk sac edema and abnormal AP axis formation (Fig. 3).

For the comparison of the morphological defects, we looked into the dose required to induce morphological defects in at least 20% of surviving embryos for each drug. The results showed that caffeine, IBMX, pentoxifyline and theophylline required ≤1 ng of drug; aminophylline and doxofylline required 3-5 ng of drug to induce the morphological defects in at least 20% of the embryos. However, diprophylline and etophylline didn’t cause morphological defects in 20% of the embryos even with the highest doses of 20 ng and 5 ng respectively.

Cardiac toxicity and Teratogenicity

The cardiovascular effects of methylxanthines in zebrafish embryos were more extensively evaluated. Zebrafish embryos were injected with the same drugs and doses (see Table 1) at 1-2 cell stage. Embryos were then analyzed for the structural and functional alterations.

Structural alteration

Assessment of the structural alteration of the heart was performed at 48 hpf. The embryos showing structural cardiac defect were termed cardiac phenotype and embryos with normal cardiac morphology were termed normal phenotype. The total percentage of embryos with cardiac phenotype was calculated for each of the compounds (Fig. 4a). Each of the embryos with cardiac phenotype was further assessed and graded based on the the criteria modified by the study of Panzica Kelly et al. [34]. The structural cardiac deformity in the embryos was scored as follows: 3 in case of normal heart morphology with slightly smaller ventricle, 2 if the atrium and ventricle were either compressed or severely enlarged and misshaped with no clear boundary between the atrium and ventricle and 1 if the atrium and ventricles were severely deficient, misshaped and not well defined.

Our results showed that with the exception of diprophylline and etophylline, all the other drugs induced cardiac phenotype in 20-40% of embryos with the highest dose. The minimum dose that caused at least 10% cardiac defects was 0.15 ng for IBMX, 0.25 ng for caffeine, pentoxifylline and theophylline,1 ng for doxofylline, 2.5 ng for aminophylline and 5 ng for etophylline. With diprophylline, the cardiac defect was present in less than 10% of the embryos even with the highest concentration of 20 ng. Further analysis of the embryos with cardiac phenotype showed that most of them had grade 2 cardiac abnormality. They had severely enlarged atrium and ventricles with distorted shape and not clearly defined border (Fig. 4c, d).

Evaluation of cardiovascular endpoints

In the embryos with structural cardiac defect, we also looked for several other cardiovascular endpoints at 48 hpf (Table 3). We observed that most of these embryos also had pericardial edema and decreased or absent blood circulation. Haemorrhage occurred only in few embryos injected with caffeine and theophylline, while thrombosis was absent in all the drug injected embryos.

Functional alteration

Heart rate in the normal looking embryos was measured at 48 hpf to assess the functional alteration of cardiac system. As methylxanthines are adenosine receptor antagonist, we expected to observe an increase in heart rate in the treated embryos if they were affected at functional level. Ten embryos were randomly selected for each concentration of each drug and the heart beat was counted. Embryos treated in sterile water were used as a negative control and metoprolol treated embryos were used as an internal control.

Our results showed that all the methylxanthines with the exception of doxofylline caused a significant increase in heart rate as compared to controls (Fig. 5). Doxofylline did not alter the heart rate whereas metoprolol significantly decreased the heart rate.

To see if the functional effect of these drugs in the embryos was a long lasting effect, the heart rate was tracked every 24 h from 48 hpf till 120 hpf in the BMP transgenic zebrafish embryos. We found that there was no significant difference in the heart rate in the treated and control embryos at 72, 96 and 120 hpf (Additional file 1).

Embryotoxicity and general Teratogenicity

In the present study, we showed that if the exposure happens during the early phase of cellular differentiation, a single dose of methylxanthines might be sufficient to induce embryotoxicity and teratogenicity. Almost all the tested methylxanthines induced embryotoxic and teratogenic effects of varying intensity. Based on their potency, these drugs could be classified into three groups: highly toxic and teratogenic, toxic and teratogenic at higher doses and non toxic and teratogenic. Caffeine, IBMX, pentoxifylline and theophylline were highly toxic and teratogenic whereas aminophylline doxofylline and etophylline induced toxicity and teratogenicity but required higher doses. On the other hand, diprophylline induced minimal teratogenicity and toxicity which was comparable to that observed in control-injected embryos. The potency of these drugs in zebrafish embryos was also similar to their potency in human and other animal models. Indeed, studies on humans and other vertebrates have shown that etophylline, doxofylline and diprophylline have lesser side effects than other methylxanthines [12, 35]. Also, in vivo and in vitro studies have shown theophylline to be more potent than etophylline (7 fold more potent), diprophylline (100 fold more potent) [35] and doxofylline (10 fold more potent) [36]. Moreover drugs like doxophylline and diprophylline have fewer side effects as compared to other methylxanthines such as theophylline [10, 12]. Overall, our results showed that a single dose of these drugs could be embryotoxic and teratogenic in the zebrafish embryos with potencies comparable to the other higher animal models.

Cardiac toxicity and Teratogenicity

The cardiac toxicity of the methylxanthines can be divided into structural and functional aspects. In our study, structural cardiac defect was not as profound as the functional effect. While the functional effect was present in all the drug treated embryos, structural cardiac defect was present in <40% of embryos at the highest concentration with all the methylxanthines. In particular, diprophylline and etophylline induced cardiac defect in less than 10% of embryos even at the highest injected concentration. Therefore, our results showed two facets of methylxanthine toxicity: structural and functional. While the structural alteration led to death, the functional alteration was transient and mild with the embryos ultimately recovering. One of the major adverse effects of methylxanthines is cardiac toxicity. In human, acute toxicity of methylxanthines cause functional alteration of cardiac physiology such as tachycardia and arrhythmia [1, 21] rather than the structural alteration of heart. However, animal studies have shown association of methylxanthines with structural cardiac defect [37, 38] and systematic review has shown cardiovascular malformations as a teratogenic effect of maternal caffeine ingestion in human [39]. Our study showed that methylxanthine induced structural alteration has lower incidence than the functional alteration. However, it could potentially be fatal. In this study, embryos were exposed to doses that fall within the therapeutic range in human. Although, interspecies differences need to be considered before drawing any conclusion, our study indicated that a single exposure of methylxanthines at therapeutic range could induce both functional and structural cardiac disfunction if exposed during early pregnancy before cellular differentiation.

Comparison of cardiac effects with higher vertebrate studies

Human and animal studies have shown that most of the methylxanthines increases the heart rate [10, 13, 15, 35, 40, 41], whereas doxofylline does not have any effect on it [10, 36, 42, 43]. These findings were similar to the results we obtained in the zebrafish embryos where all the methylxanthines increased the heart rate and doxofylline did not have any effect. Therefore, at functional level, there was similarity in the effects of methylxanthines in zebrafish embryos and other animal models and human. The ability to respond to both the cardiac stimulant such as methylxanthines and depressant drug such as metoprolol also confirmed the functional maturity of the cardiac system of zebrafish embryos at 48 hpf, which is well supported by other zebrafish studies (mentioned in Review [20, 44]).