Regulatory Forum Opinion Piece*

Abstract

We recently conducted a retrospective analysis of data collected from 29 Tg.rasH2 carcinogenicity studies conducted at our facility to determine how successful was the strategy of choosing the high dose of the 26-week studies based on an estimated maximum tolerated dose (MTD). As a result of our publication, 2 counterviews were expressed. Both counterviews illustrate very valid points in their interpretation of our data. In this article, we would like to highlight clarifications based on several points and issues they have raised in their papers, namely, the dose-level selection, determining if MTD was exceeded in 26-week studies, and a discussion on the number of dose groups to be used in the studies.

We recently conducted a retrospective analysis of data collected from 29 Tg.rasH2 carcinogenicity studies conducted at our facility to determine how successful was the strategy of choosing the high dose of the 26-week studies based on an estimated maximum tolerated dose ([EMTD]; Paranjpe et al. 2015). To analyze whether or not the high dose in the 26-week studies exceeded the maximum tolerated dose (MTD), we took into consideration the initial body weights (IBWs), body weight gains (BWGs), food consumption (FC), mortality, and percentage of tumor incidence in each of the dose groups (control, low, mid, and high). Further, we collected data on the cause of death (COD) in each of the groups, which we consider to be a very important parameter for determining if the death was related to treatment and/or caused by related neoplastic or nonneoplastic lesions, or was the COD undetermined, which simply indicates excessive toxicity.

We demonstrated that while the IBWs and FC did not differ between any of the groups, the high-dose groups in both sexes caused a greater than 10% drop in BWGs and a significant increase in mortality. Also, the COD due to tumors was highest in the control groups of both sexes and it gradually decreased in a dose-responsive manner from high- to low-dose groups. In addition, COD classified as undetermined was highest in the high-dose group of both sexes. We clearly demonstrated that decreases in BWGs and increasing mortality, not related to tumor formation, affected tumor formation in the high-dose groups. By comparison, the tumor incidence increased in low- and mid-dose groups.

We value and appreciate the opinions expressed by Nambiar and Morton (2015) and Darbes, Sistare, and DeGeorge (2015) in their counterpoints. We are grateful to all the authors for their responses to our recent publication. Both papers illustrate very valid points in their interpretation of our data, but we would like to highlight some clarifications based on several points and issues they have raised.

Dose-level Selection for the 29, 26-week Studies Conducted by BioReliance

One issue that was raised by both counterpoints is the lack of clarity on how dose levels were chosen for our studies. Although we did not provide a breakdown of how dose levels were chosen, for more than 90% of the 26-week studies in our database, the high dose was the EMTD, as estimated based on data from 28-day range finding studies, conducted using the wild-type littermates of the Tg.rasH2 mice. As discussed by us and both counterpoints, the criteria used for estimating the MTD included test article–induced mortality, >10% lower BWG relative to the concurrent controls, and significant target organ toxicity.

In determining the EMTD based on 1-month dose range finding studies, we clearly stated the parameters that were considered included—mortality, BWG, tissue toxicity, clinical observations, and clinical pathology findings. The dose levels for each of the 26-week studies we reported were approved/selected by the Carcinogenicity Assessment Committee (CAC) at the U.S. Food and Drug Administration (FDA), based on the results of the 28-day studies. We always recommend that the CAC committee reviews and approves all 26-week study designs to ensure that dose levels are selected based on the applicable guidelines and to ensure consistency in dose selection across all studies. We do not make independent decisions on the protocols and the selection of doses without involving the sponsors and obtaining approval from the Executive CAC (ECAC)/CAC at FDA.

For our studies, mortality was used as the main end point in selecting dose levels in most of the 26-week studies we reported. Up until our recent publication, BWG was not critically analyzed and considered in determining the dose levels for the 26-week studies. Body weight changes in 28-day studies that were lower (<10%) and/or not lower in a statistically significant manner were often disregarded. In addition, there have been some instances for which dose levels selected for the 26-week studies had caused some tissue toxicity and/or clinical signs suggestive of systemic toxicity that did not result in mortality or significant changes in BWG in the 28-day studies. As a result, these dose levels caused either lower BWGs and/or mortality in the 26-week studies.

It is important to keep in mind that dose levels chosen for the 26-week studies are chosen based on dose range findings that are only 28 days in duration. Often, unless the test article is known to cause lower body weights or the dose levels chosen are much higher than the EMTD, body weight changes in the 28-day study will be minimal, even in the presence of excessive toxicity. In addition, slight changes in BWGs (<10% and/or not lower in a statistically significant manner) in the 28-day studies are likely to translate to more significant changes when the same dose is dosed over a 26-week period, especially since the Tg.rasH2 mice are smaller in size than the wild-type littermates used in the range finding studies.

We agree with Darbes, Sistare, and DeGeorge (2015) in that sponsors tend to suggest high doses for the carcinogenicity studies for multiple reasons. Our experience has also been that CAC will tend to recommend lowering the dose levels to minimize the risk of excessive toxicity. We have never been questioned on dose selection for a 26-week study by any regulatory agency after the studies have been conducted.

Determining if MTD Was Exceeded in the 26-week Studies

After analyzing the data reported by Nambiar and Morton (2015) and Darbes, Sistare, and DeGeorge (2015), we feel it is necessary to clarify how we determined that MTD was exceeded in the 26-week studies. The criteria used by us to qualify the number of studies that exceeded the MTD were not solely based on >10% decrease in BWG relative to control. In fact, that was only one of the factors. The other factors considered were mortality and the incidence of tumors in all dose groups. Using the well-accepted definition of MTD, it is the dose that causes only minimal toxicity, a reduction in BWG that is not more than 10% of the BWG in controls, and/or does not increase mortality or alter the longevity of the animals due to reasons other than tumors. In fact, we consider our analyses of COD as a very important parameter in determining the MTD. Our analyses clearly showed that in addition to the lower BWG and increased mortality, the mortality was not due to increased incidence of tumors. In fact, mortality due to tumors was highest in the control groups of both sexes and lowest in the high-dose groups. Our analyses also showed that undetermined, as the COD, was highest in the high-dose groups of both sexes. Thus, it is clear that the animals in the high-dose groups did not die because of tumors, but they died because of toxicity, indicating that the doses exceeded the MTD.

We do not think that the data we presented differ significantly than those published by Darbes, Sistare, and DeGeorge (2015) or by Nambiar and Morton (2015). Nambiar and Morton indicate that 8 of 11 studies used MTD as the rationale for selecting the high dose in the 26-week study. For the remaining 3, 1 used maximum feasible dose (MFD), 1 used limit dose, and 1 high dose was based on a BWG decrement greater than 20% relative to controls. However, by definition, the third study using BWG decrement, was in fact based on the MTD since a BWG decrement >10% relative to control is used to determine an MTD. Thus, 9 of the 11 studies reported by Nambiar and Morton used MTD to identify dose levels for the 26-week carcinogenicity studies. The authors also state that in all but one study, the high dose was well tolerated. Although based on the limited information available, it does appear that mortality was not altered in any of the studies; it is apparent that there were body weight changes in many of the studies reported for the various compounds. For the 9 compounds for which the MTD was used as the criteria to determine dose levels (changes in BWG and/or increase in mortality), 2 had a difference in body weight (BW) that was −10% to −20% compared to control and 1 had more than −20% difference. In addition, for one study (compound K), the high dose that was initially requested by FDA (75 mg/kg/day) caused early mortality and was reduced during the conduct of the study. Thus, by our account, 4 (∼44%) of the 9 studies reported by Nambiar and Morton that used MTD as the criteria for setting dose levels had a dose that exceeded the MTD based on the results of the 26-week study. In addition, the authors only reported the difference in absolute BW versus controls in the 26-week studies, and not the difference in BWG, which is the criteria used for assessing the MTD. From our experience, differences in BWGs between the controls and treatment groups do not always translate to similar differences in absolute BWs. Thus, for some of the studies for which no change in BW versus control is reported, some may in fact have a difference in BWG. As an example, for compound E, there was a decrease in BWG in males at 2,000 mg/kg/day in the range finding study, yet the same dose did not cause a change in absolute BW in the 26-week study. Thus, for the studies reported by Nambiar and Morton, at least 44% had a dose that exceeded the MTD by definition.

Similarly, in the paper by Darbes, Sistare, and DeGeorge (2015), for 6 studies (compounds A, B, C, D, E, and F) of the 11 presented, the high dose was selected based on estimation of the MTD. For one of those studies (compound B), the high dose was pushed high, close to the lethal dose noted in the range finding study and was subsequently not tolerated in the 26-week study. Thus, in looking at how accurate the MTD was estimated in the remaining 5 studies, 3 of the 5 had high-dose levels that exceeded the MTD based on the most well-accepted definition of MTD (Paranjpe et al. 2015), which specifies more than 10% decrease in BWG relative to control and/or increased mortality not related to tumor formation. The high dose for compound C caused a −61% to −80% decrease in BWG; there was 48% mortality in the females at the high dose for compound E; and the high dose for compound F had to be reduced and then terminated due to increased mortality and decreased BWGs. Our analysis did not focus on tolerability of the high dose in the 26-week studies based on survival but rather whether or not it was truly an MTD based on body weight effects and mortality. The results shown by Darbes, Sistare, and DeGeorge (2015) confirm our findings that the MTD was not accurately predicted, and was exceeded, in more than 50% of the 26-week studies.

It would have been beneficial if both papers referenced above provided statistics on tumor incidence in each of the groups. While most studies were negative for compound-related tumor induction, there is no indication of a trend in tumor incidence based on differences in BWGs between the various groups, similar to what we have presented in our article.

We agree that it would require huge differences in BWGs to produce noticeable differences in mean body weight at the end of the 26-week studies. However, our results clearly show a significant trend toward lower BWGs in the high-dose groups in most 26-week studies that have been conducted. While this difference may be minimal in terms of body weight differences, it caused a significant impact on tumor formation at these high doses but not at the low- and mid-dose levels. While we also agree there is large spontaneous variability in BWG over the course of a 28-day study, we have found that looking at trends in body weight changes and correlating them with any histopathology findings or even clinical signs of toxicity/exaggerated pharmacology helps in identifying toxicologically significant changes in body weight. We would also like to emphasize that body weight effects may be masked in range finding studies. For example, compounds that cause excessive edema or enlargement of tissues (i.e., liver hypertrophy) may cause significant target organ toxicity or clinical signs with no adverse effect on BWGs. BWGs are one tool we use to identify the MTD, but we always consider other toxicologically relevant end points as well.

The Number of Dose Groups

Both Darbes, Sistare, and DeGeorge (2015) and Nambiar and Morton (2015) raise valid points about the importance of the three dose groups, in addition to negative/vehicle control and positive control groups. All of our studies include a vehicle (and sometimes a negative) control, in addition to the positive control group and three test article–treated groups. In our article, we suggested that studies can be performed with 2 dose groups, instead of 3, if the dose levels are chosen appropriately. However, we also stated that, if needed a third dose group may be added. The decision pertaining to the number of dose groups has to be made on a case-by-case basis. Darbes, Sistare, and DeGeorge (2015) state that when the spread in a meaningful dose selection is narrow due to low human exposure multiples being achievable, it may seem more reasonable to use only 2 dose levels to conserve on animal use. Some of the other scenarios that may allow for a 2-dose-level study include (a) if abundant mechanistic information is available pertaining to the test article and it is not expected to be a carcinogen or (b) if prior 28-day wild-type mouse, 6-month rat, or 2-year rat studies have not shown any toxicity/carcinogenicity

We agree that there is always a residual risk for exaggerated mortality or significant toxicity, despite adequate dose selection. We fully understand the requirement of three dose groups in a carcinogenicity assay and that the three dose groups may actually be absolutely necessary to draw the No Observed Effect Level/No Observed Adverse Effect Level (NOEL/NOAEL) and/or dose-dependent increase in the tumors, if any. However, our argument is mainly against the high-dose group at EMTD, as it is currently chosen, which did not serve the purpose at least in our assays. The combination of decreased BWG, increased mortality that was not due to tumor formation, and the fact that undetermined as the COD was highest in the high-dose groups and lowest in the control dose groups were factors we used to determine if MTD exceeded. So, if the high-dose group at EMTD is not going to serve the purpose in a carcinogenicity assay, which is to detect the carcinogenic potential of the test article then that dose group needs to be eliminated and replaced by lower doses such as 1/2 of EMTD in males and 2/3 of EMTD in females derived from 28-day studies, as we have suggested in our article. If these high doses are lowered as suggested by us, then the risk of exaggerated mortality and significant toxicity will be reduced. These 2 latter points are also indirectly pointed out by both Nambiar and Morton (2015) and Darbes, Sistare, and DeGeorge (2015) wherein they found that MTD exceeded in approximately 44% and 50% of the studies, solely based on changes in BWG. However, neither the study by Nambiar and Morton (2015) nor the study by Darbes, Sistare, and DeGeorge (2015) discuss the incidence of tumors in individual dose groups, nor do they discuss the COD. Our main take-home message is that if the EMTD derived from current applicable criteria based on 28-day studies is not going to serve the purpose in the 26-week carcinogenicity assay, then high-dose selection in the 26-week Tg.rasH2 studies has to be modified.

Our aim is to reduce the likelihood of using a top dose that exceeds the MTD so that only the slight, residual risk remains a factor in a top dose causing excessive toxicity and/or body weight changes that may alter the tumor outcome in that group. The purpose of our article was mainly to highlight some deficiencies in selecting the top dose level, specifically using the EMTD, for carcinogenicity studies using Tg.rasH2 mice. We do acknowledge that saturation of systemic exposure and maximum feasible dose is still used in determining the dose levels. However, from our experience, these 2 criteria were rarely used. Not many compounds show saturation of exposure before signs of toxicity are demonstrated. The limit dose can only be used as criteria when various formulations have been tested. Having exposure margins up to 25× the area under the curve (AUC) of the human dose is not always achievable in rodents, simply because rodents do not always tolerate compounds the same way humans do. ECAC has also acknowledged that exposure margins are not useful in determining dose levels for 26-week studies. In the end, we are only trying to refine the process by which the EMTD is selected in the 26-week studies.

Conclusion

We agree with the concept of using EMTD, based on data from the 1-month range finding study, as the high dose for the 26-week Tg.rasH2 studies. However, we recommend that the MTD continues to be the primary end point for choosing the high dose, in the absence of saturation of the toxicokinetic (TK) profile, and when feasible. In our article, we simply wanted to illustrate how ineffective the current process has been at predicting the MTD since this model has been in use, which has also been indirectly suggested by both counterreviews, by Nambiar and Morton (2015) and Darbes, Sistare, and DeGeorge (2015).

As Darbes, Sistare, and DeGeorge (2015) have stated, it would be of significant benefit to add dose selection criteria for the 26-week studies in the International Conference on Harmonisation, Dose selection for carcinogenicity studies (ICH S1C) guidance, particularly in light of the data we, and others, have provided.