IMMUNOLOGY OF DIABETES V

What Will It Take to Get Therapies Approvedfor Type 1 Diabetes?

Alexander Fleming

Kinexum, Harper’s Ferry, West Virginia, USA

The development of therapies for T1D has been neglected in favor of efforts in advancingtherapies for the larger T2D population. Pharmaceutical companies have also beendeterred by lack of clarity around the regulatory expectations for such therapies. Theprospects for therapy for new-onset T1D have brightened in some respects because ofconvergence among regulators and clinical experts in views about how these therapiesshould be assessed. The most important consensus is that the primary efficacy endpoint for treatments directed at the underlying autoimmune cause of T1D should beendogenous insulin secretion, as reflected by standardized C-peptide measurements.Most T1D therapeutic development efforts are directed at new-onset disease, whichrepresents a small proportion of the entire T1D population. A major deficiency inT1D therapeutic development is the lack of activity in advancing therapies for peoplewith established T1D, a population that far outnumbers those with new-onset disease.Complete remission of new-onset or established T1D will almost certainly require acombination of two or more therapies to address the underlying cause of the diseaseand restore normal insulin function.

Key words: FDA; T1D; immunomodulatory; regulatory; C-peptide; insulin; hypo-glycemia

Introduction

The United States Food and Drug Admin-istration (FDA) and other regulatory authori-ties have long been in a quandary about howto approve therapies directed at the under-lying autoimmune cause of type 1 diabetes(T1D). Therapeutic targeting of islet autoim-munity has been focused to date on personswith some remaining β-cell function, essentiallythose just diagnosed with T1D. This target pop-ulation requires reenrolling in a registered clin-ical trial, which are relatively long and difficultto conduct. These and other challenges haveslowed the pharmaceutical industry’s pursuit

Address for correspondence: Alexander Fleming, Box 1260, 550 RidgeStreet, Harper’s Ferry, WV 25425. Voice: +304-535-3037; fax: +304-535-3166. zanfleming@kinexum.com

of T1D therapies, even though it afflicts morethan 1 million North Americans.1 The largertype 2 diabetes (T2D) population has domi-nated the attention of regulators and pharma-ceutical companies alike. For T2D drugs, theregulatory efficacy end point, hemoglobin A1c(HbA1c), was validated by landmark trials aspredictive of microvascular benefit,2,3 and thedesign of T2D registration trials has becomestandard. For some time, the FDA assumed thatHbA1c should also be the primary end point forT1D therapies. However, this position causedconcern among the expert community. Clin-icians and trial designers understood that forethical and analytical reasons, T1D trials haveto consist of a treat-to-glycemic-target design.This design will tend to minimize treatmentgroup differences in HbA1c. Glycemic controlwould, therefore, be an insensitive measure ofexpected clinical benefits for clinical trials inT1D.

Immunology of Diabetes V: Ann. N.Y. Acad. Sci. 1150: 25–31 (2008).doi: 10.1196/annals.1447.043 C© 2008 New York Academy of Sciences.

25

26 Annals of the New York Academy of Sciences

Recent History

This initial regulatory policy led to an Oc-tober 2001 workshop of international experts,which was sponsored by the American DiabetesAssociation (ADA) and co-sponsored by the Na-tional Institutes of Health (NIH) and the Ju-venile Diabetes Research Foundation (JDRF).A consensus was reached that the appropri-ate primary efficacy outcome for T1D inter-vention trials is endogenous insulin secretionas measured by C-peptide levels.4 Around thistime, the European Agency for the Evaluationof Medicinal Products (EMEA) issued a guide-line for T1D and T2D therapies, but it did notaddress therapies for the underlying cause ofT1D. However, sponsors who conferred withEuropean national authorities were generallygiven advice consistent with the ADA consen-sus.

The publication from the ADA workshop in2003 containing an analysis of the DiabetesControl and Complications Trial (DDCT) dataproved ultimately persuasive to the FDA. Thisanalysis examined the relationship of endoge-nous insulin secretion at study entry (>1 yearafter diagnosis of T1D) and rates of microvascu-lar complications and glycemic control at long-term followup. The analysis suggested thata relatively small preservation of endogenousinsulin secretion, as reflected by C-peptide lev-els, is associated with improved microvascu-lar and glycemic outcomes irrespective of thestudy treatment assignment, including the in-tensive glycemic control group. The results ofthe analysis and the expert consensus supporteda change in the FDA’s view about the pri-mary efficacy endpoint for T1D clinical tri-als. Though the analysis of C-peptide datawas not prespecified, the use of C-peptide asa more appropriate primary efficacy end pointfor T1D therapeutic development is now widelyaccepted.

This article discusses impressions of the con-verging perspectives of the FDA, EMEA, andthe expert community.

Target Populations and TherapeuticIndications

The FDA and EMEA both recognize thecurrent unmet clinical need for a new meansof preserving remaining endogenous insulin se-cretion in patients with new-onset T1D. Bothagencies would readily approve safe and effec-tive treatments for the retention and/or restora-tion of endogenous insulin secretion, as wellas agents for the prevention of T1D; however,there are a number of key considerations re-garding the determination of the target popu-lations of such treatments.

Initial pivotal trials involving immunomodu-lators will generally be limited by regulators toadults and older children; the indicated popu-lation would then be defined by the parametersused in the pivotal trials. After the establish-ment of safety and efficacy in early trials, ad-ditional trials including progressively youngerchildren will be allowed, and eventuallyrequired.

Until actual clinical benefits are demon-strated, the therapeutic indication would pre-sumably specify that the treatment is for thepreservation of remaining endogenous insulinsecretion in people with recently diagnosedT1D. It is anticipated that some limited men-tion of the DCCT analysis would be allowedin the drug product label, but no claims couldbe made for expected clinical outcomes, suchas the prevention of complications.

Study Design

Duration and Durability

The FDA and EMEA now approve thera-pies for T2D based on pivotal trials that are6 months in duration or longer. For T1D ther-apies, it is likely that the FDA will require atleast 12 months of controlled trial treatmentwith indefinite follow-up of the study cohorts.However, concerns about safety and durability

Fleming: Regulatory Expectations for Developing T1DM Therapies 27

of effect may drive trials of longer duration.Durability can be evaluated during the courseof the controlled trial and, to some extent, dur-ing the unblinded follow-up.

All T1D studies must adhere to acceptedstandards of care in managing glycemia andco-morbidities. In particular, trial partici-pants should be managed to achieve near-normal glycemic control. Incorporating ag-gressive glycemic targets in T1D trials will alsoimprove the ability to demonstrate effects onthe secondary efficacy outcomes—rates of hy-poglycemia and total daily insulin dose.

Efficacy End Points

Primary End Point: Endogenous InsulinSecretion as Reflected by StimulatedC-Peptide Levels

T1D results in a hormonal deficiency state.Clearly, the appropriate primary basis for as-sessing the efficacy of a T1D therapy is theextent to which this deficiency state is cor-rected. As discussed above, clinical benefitscan be expected from preserved autoregu-lated insulin secretion. C-peptide is often re-ferred to as a surrogate efficacy outcome, butit is important to understand that C-peptideis simply a more reliable means of measur-ing insulin secretion than measuring insulinitself. C-peptide therefore is better viewed asa part of a methodologic approach for assess-ing endogenous insulin secretion rather than asurrogate outcome or biomarker. Endogenousinsulin secretion should be assessed by mea-surement of C-peptide levels under standard-ized conditions—currently a mixed-meal toler-ance test (MMTT) or glucagon stimulation testare commonly used. An international study hascompared these two approaches and the resultsare expected soon.5

Secondary End Points and TheirConsiderations

HbA1c is a highly valuable clinical measureof glycemic control, but it is an insensitive mea-

sure of β-cell function, particularly in a trial inwhich near-normal glycemic control is sought.In such trials, HbA1c is not expected to sub-stantially differ between treatment groups; infact, a trial that results in substantial differencesin HbA1c between treatment groups would beproblematic from the standpoint of interpretingthe results since glycemic control conceivablymay affect residual β-cell function. Therefore,HbA1c is regarded as an important measureof trial integrity and as a check against adverseeffects of the therapy on glycemic control.

Total daily insulin dose is another impor-tant efficacy end point, but has significantlimitations. The insulin dose is expected to dif-fer between placebo and treatment group. In-sulin dose is also influenced by multiple factors(e.g., exercise, type of insulin used, intra-individual insulin pharmacokinetics and dy-namics, insulin sensitivity). Thus, this param-eter is likely to have more variability than theunderlying endogenous insulin secretion. Froma biostatistical perspective, it would be chal-lenging to adequately power a trial with ameasure of much higher variability than thatof the endogenous insulin secretion. In addi-tion, no generally accepted magnitude of effecton insulin dose that could be considered clin-ically meaningful or predictive of clinical ben-efits has been established. Therefore, insulindose will be seen as a supportive secondary endpoint.

Hypoglycemia is the most important bar-rier to achieving normal glycemic control inpeople with T1D. Preservation of endogenous,glucose-regulated insulin secretion shouldreduce the rates of serious hypoglycemia as-sociated with insulin therapy. Reduction of hy-poglycemia is therefore the raison d’etre of ther-apies aimed at the underlying cause of T1D.Accordingly, the FDA views the assessment oftreatment effects on rates of hypoglycemia asvery important, but is also cognizant of thechallenges in assessing this outcome, especiallyover the relatively short duration of preapprovalclinical trials. The event rate of hypoglycemia

28 Annals of the New York Academy of Sciences

as rigorously defined in the DCCT is low, anda clear benefit on hypoglycemia is not expectedin the first years after diagnosis of T1D. Like-wise, the benefit of intensive glycemic controlon rates of hypoglycemia in the DCCT wasnot seen until after the third year of that study.Hypoglycemia can be assessed in various ways,including the use of continuous glucose mon-itoring, multiple home glucose monitoring us-ing a standardized instrument, and self- andfamily-reported hypoglycemia, but regulatorswill continue to rely on the rigorous criteriaestablished in the DCCT.

Statistical Approaches and Effect Size

Standard statistical approaches are to beused to size the pivotal T1D trials and provideprespecified analysis plans. The key unresolvedquestion is what would be regarded by the reg-ulatory authorities as the minimally acceptabletreatment effect size. Given the lack of availabil-ity of an approved therapy, it is conceivable thatFDA would accept on the order of a 20–30%placebo-adjusted difference in C-peptide secre-tion. A categorical analysis approach might alsobe accepted. This might consist of the differ-ence in the proportion of subjects maintain-ing a C-peptide response above a prespecifiedlevel.

Basis for Regulatory Approval

As is true for the licensing of any therapeu-tic product, the regulator weighs the observedor likely benefits against the risks suggested byclinical and nonclinical data to reach a deci-sion about approvability and the appropriatedrug product label. The FDA and EMEA nowunderstand that some partial preservation ofendogenous insulin secretion should result inclinical benefits. The expert community’s ad-vice will be sought on what is a minimally ac-ceptable treatment effect and just how mini-mally durable that effect should be. While sucha threshold may become an established stan-

dard, the magnitude of benefit that would berequired for a particular therapy would dependon that specific therapy’s safety profile.

Islet Regeneration Therapies andCombination Approaches

The large majority of persons with T1D can-not expect to benefit from immunomodulatorytherapy alone since it does not directly ad-dress the almost total loss of β-cell mass, whichotherwise occurs within a few weeks of clin-ical onset. Evidence is building that progen-itor or stem cells reside in the ductal tissueof the pancreas, even after many years of dis-ease, and may be induced to differentiate intofunctional islets,6 but this remains an area ofcontroversy.7 A number of therapeutic candi-dates have shown true islet-regenerating activ-ity (to be distinguished from agents that mayinduce β-cell replication or reduce β-cell apop-tosis) in animal models and at least two areunder clinical development.8,9 Figure 1 exem-plifies the value of combining a therapy thatcontrols the autoimmune destruction of isletswith a therapy that induces islet neogenesis.In the NOD mouse model of well-establishedT1D, an immunomodulatory agent, liso-fylline,10 and an islet-regenerating agent, IN-GAP (islet neogenesis-associated protein) pep-tide.11 When given alone lisofylline resulted inno reduction in hyperglycemia. INGAP aloneresulted in only modest reduction of hyper-glycemia, although insulin levels increased sub-stantially. When both therapies were giventogether, insulin levels normalized and hyper-glycemia was reversed. Figure 1 shows appar-ent regeneration of islets as evidenced by insulinand PDX staining resulting from the combinedtreatment.12

Combination therapeutic approaches con-ceivably can extend the potential for restorationof normal metabolic control by returning notonly endogenous insulin secretion, but alsosecretion of other important islet hormonessuch as glucagon for all people with T1D.

Fleming: Regulatory Expectations for Developing T1DM Therapies 29

Figure 1. Regenerated islets with insulin and PDX staining resulting from combined treat-ment with INGAP and LSF in NOD mice with established T1D. The result of combined islet re-generation and immunomodulatory therapy is demonstrated with immunohistochemical stain-ing for insulin (left panels) and the islet marker, PDX-1 (right panels). NOD mice were startedon insulin pellets after three successive measures of plasma glucose exceeding 250 mg%. LSFwas then started followed by INGAP peptide treatment. Vehicle- and monotherapy-treatedanimals showed essentially no clinical response and minimal histologic evidence of new isletformation (data not shown). (Illustration courtesy of Dr. Jerry Nadler, Jeffrey Carter, SarahTersey, and Elizabeth Kropf.)

Since those with established T1D start withessentially no endogenous insulin secretion,treatment effects resulting from islet regener-ation as measured by C-peptide as well as sup-portive secondary endpoints will require fewersubjects and less time to detect. The FDAhas well-established principles for developingcombination therapies. Generally a toxicologystudy of the combined therapy is required in ad-dition to the complete package of nonclinicalstudies required for each individual therapy.13

At least one clinical study must be conductedthat demonstrates in a factorial design the con-tribution of each agent to the combined treat-ment effect.14

Substantial efforts aimed at developing drugcombinations in the T2D therapeutic area arethe rule and not the exception. However, large

pharmaceutical companies and investors havetraditionally shied away from co-developmentof two combination therapies, one of which hasnot already been approved. This is no longerthe case, particularly in areas that have been re-sistant to monotherapy approaches, obesity be-ing probably the most relevant example.15 A re-cent example of a co-developed investigationalcombination drug product approved by thesame FDA division that regulates diabetes prod-ucts is Mecasermin rinfabate (iPlex, Insmed,Inc.), which consists of insulin-like growth fac-tor (IGF-1) and its plasma binding protein,IGF-binding protein-3 (IGFBP-3).16 Availabil-ity of robust treatments to people with T1Dwill be substantially delayed if investigationalagents are only developed in isolation and insequence.

30 Annals of the New York Academy of Sciences

Other Considerations

Mechanism of Action—VaccineParadigm

Most T1D therapies for T1D autoimmu-nity will essentially function as vaccines, notas drugs, and the development of vaccines anddrugs differ in substantial ways. For example,clear dose–response relationships are not typi-cally seen with vaccines as they are with clas-sical pharmacologic agents. The FDA can beexpected to apply a mix of drug and vaccineperspectives in evaluating immunomodulatoryapproaches for both efficacy and safety.

Dose and Regimen Optimization

Because T1D therapies will typically not beamenable to exploration of multiple doses andregimens in clinical trials, animal model datamay help to support going forward with a sin-gle dose and regimen in pivotal trials. The FDAmay require some assessment of how long achronic or recurrent therapy should be con-tinued or repeated. Randomizing trial subjectsto various regimens, including no therapy af-ter the completion of the initial treatment pe-riod, would provide valuable additional data.Presumably, this portion of the trial would becompleted after initial approval.

Durability of Effect

In the case of T2D, durability of effect has notbeen formally evaluated, though it is clearly animportant aspect of a therapy’s performance.Durability, or lack thereof, of a T1D therapywill be taken into the overall benefit-to-risk as-sessment (see below) made in the initial drug ap-proval process. That is, the FDA does not likelyhave specific expectations for demonstratingdurability; however, the FDA is likely to requiresome formal assessment of long-term durabil-ity, if not clinical outcomes, in the postapprovalperiod.

Regulatory Provisions

New-onset T1D, depending on the defini-tion, qualifies for Orphan Drug status at the

FDA and EMEA. It appears that the FDA isnot likely to grant “fast track” status routinelyfor T1D.

Public Health Policy and General DrugDevelopment Issues

Drug safety issues such as those involvingVioxx (rofecoxib) and Avandia (rosiglitazone)have put the FDA under public and Congres-sional scrutiny. While these concerns generallycenter on therapies for large populations, theimpact at the FDA is to cause greater risk aver-sion among reviewers. This effect may extendinto review of therapeutic efforts aimed at clearunmet clinical need, such as those for T1D. Asomewhat related challenge for the pharma-ceutical industry is the increasing costs of ther-apeutic development. Drug development is in-herently risky and costly, mainly due to attritionof drug candidates from company pipelines.The cost of producing a single approved newmolecular entity approaches $1 billion.17 Themost costly losses arise from late-stage cancel-lation of development programs, which arisefrom unforeseen toxicity issues or from lack ofenough clinical benefit to outweigh risks of ad-verse events. Though the FDA and EMEA mayaccept a somewhat smaller number of totalpatients in T1D development programs, thelong duration of these trials may lead to totalcosts that approach those of T2D drugs. En-couragingly, some T1D therapeutic programshave attracted major pharmaceutical companyinvolvement,18 but development of all diabetestherapies and therapies for T1D in particularwill be valued by industry with consideration ofthese economic and political pressures.

Conclusions

The prospects for new-onset T1D therapeu-tic development have improved, in part, be-cause of the convergence of regulators’ andclinical development experts’ thinking abouthow these therapies should be assessed forefficacy. A number of key issues remain to be

Fleming: Regulatory Expectations for Developing T1DM Therapies 31

resolved. These will be faced soon since sev-eral immunomodulatory candidates have ap-proached or entered Phase 3 clinical trials.For persons with established T1D, a popula-tion that far outnumbers those with new-onsetdiabetes, a major deficiency in development ef-forts remains. Complete remission of all cases ofT1D will likely require a combination of two ormore therapies, in a sense, similar to therapeu-tic combinations required for metabolic controlof T2D. While challenges remain, noninvasivemedical therapies to prevent and reverse au-toimmune destruction of islets are no longervain hopes.

Acknowledgments

The material shown in Figure 1 was the resultof research funded by the Iacocca, Farish, andElla Fitzgerald Foundations and by the Ameri-can Diabetes Association.

Conflicts of Interest

Serve as a paid advisor to many companiesdeveloping T1 and T2 therapies. I hold a majorinterest in the development of a islet neogensiscompound.

References

1. National Institute of Diabetes and Digestive and Kid-ney Diseases. 2003. National Diabetes Statistics Fact Sheet:

General Information and National Estimates on Diabetes in

the United States, 2003. U.S. Department of Healthand Human Services, National Institutes of Health.Bethesda, MD.

2. The DCCT Research Group. 1993. The effect ofintensive diabetes treatment on the development andprogression of long-term complications in insulin-dependent diabetes mellitus. N. Engl. J. Med. 329:978–986.

3. The UKPDS Study Group. 1998. Intensive blood-glucose control with sulphonylureas or insulin com-pared with conventional treatment and risk of com-

plications in patients with type 2 diabetes (UKPDS33). Lancet 352: 837–853.

4. Palmer, J.P., G.A. Fleming, C.J. Greenbaum, et al.2004. C-peptide is the appropriate outcome measurefor type 1 diabetes clinical trials to preserve beta-cellfunction: report of an ADA workshop, 21–22 October2001. Diabetes 53: 250–264.

5. TrialNet Investigators. Improving Metabolic Assess-ments in Type 1 Diabetes Mellitus Clinical Trials.http://clinicaltrials.gov/show/NCT00105352.

6. Bonner-Weir, S.B. & A. Sharma. 2006. Are therepancreatic progenitor cells from which new isletsform after birth? Nat. Clin. Pract. Endocrinol. Metab.

2: 240–241.7. Dor, Y. 2006. β-cell proliferation is the major source

of new pancreatic cells. Nat. Clin. Pract. Endocrinol.

Metab. 2: 242–243.8. Suarez-Pinzon, W.L., Y. Yan, R. Power, et al. 2005.

Combination therapy with epidermal growth factorand gastrin increases β-cell mass and reverses hyper-glycemia in diabetic NOD mice. Diabetes 54: 2596–2601.

9. Fleming, A. & L. Rosenberg. 2007. Prospects andchallenges for islet regeneration as a treatment fordiabetes: A review of islet neogenesis associated pro-tein. J. Diab. Sci. Technol. 1: 231–244.

10. Yang, Z., M. Chen, J.D. Carter, et al. 2006. Com-bined treatment with lisofylline and exendin-4 re-verses autoimmune diabetes. Biochem. Biophys. Res.

Commun. 344: 1017–1022.11. Rosenberg, L., M. Lipsett, J.W. Yoon, et al. 2004.

A pentadecapeptide fragment of islet neogenesis-associated protein increases beta-cell mass and re-verses diabetes in C57BL/6J mice. Ann. Surg. 240:875–884.

12. Nadler, J.L. Manuscript in preparation. Quoted bypermission.

13. Guidance for Industry. Nonclinical Safety Evaluation of

Drug Biologic Combinations. U.S. Department of Healthand Human Services, Food and Drug Administra-tion, CDER. March 2006. www.fda.gov

14. Food and Drug Administration 21 CFR 300.50.15. Morrison, T. Combo drugs prepare to tap un-

derserved obesity market. Online at www.bioworld.com/.

16. www.insmed.com.17. DiMasi, J., R. Hansen & H. Grabowski. (2003). The

price of innovation: New estimates of drug develop-ment costs. J. Health Econ. 22(2): 151–185.

18. News in Brief. 2007. Selected research collaborations.Nat. Biotechnol. 25: 1339–1340.