Archives of
The National Diabetes
Center
May, 1993
Reload The National Diabetes Center
Frames
Home Page
Perspectives in Clinical Diabetes, Obesity, and
Atherosclerosis: Type 2 Diabetes as "Glucose
Insensitivity"
The "Stuffed-Cell" (c) Hypothesis, the
"Shifted Glucose-ATP/ ATP-Insulin Dose Response Curves"
Hypothesis, and the "Gluconeogenic Suppression Fallacy"
- Putative Relationships to Genetically Derived ß-Cell and
Target Organ Susceptibilities to Glucose Toxicity
Ronald J. Innerfield, M.D.-
-- Knowledge cannot spring from experience alone, but only
from the comparisons of the inventions of the intellect with the
observed fact. -Albert Einstein
After basking for some time in
the intense light of medical knowledge, our serpentine concepts
of disease tend to exchange their old external habiliments for
newer and [hopefully] more compatible ones. Thus "Uremic
Poisoning" has been shed for "Bright's Disease of the
Kidneys" which in turn has molted to "Chronic
Pyelonephritis" and "Chronic Glomerulonephritis".
On the one hand, "Chronic Pyelonephritis" has been
considered the end-stage of more acute "Interstitial
Nephritides" including "Analgesic Abuse
Nephropathy" and toxicities secondary to other pharmacologic
agents as well as secondary to infectious agents , metabolic
deficiencies, or vascular insufficiencies. "Chronic
Glomerulonephritis", on the other hand, is currently
perceived as a final common pathway of expression for immunologic
(Systemic Lupus Erythematosus, Acute Diffuse Post-Streptococcal
Nephritis), infiltrative (Amyloidosis, either primary or
secondary to Rheumatoid Arthritis or Tuberculosis), metabolic
(Diabetic Glomerulosclerosis or Kimmelstiel-Wilson Disease), or
idiopathic disorders (Membranous Glomerulonephritis,
Minimal-Change Nephropathy).
Although viewed by astute
clinicians (Himsworth, H., Lancet 1:127, 1936\;Farber SJ,
personal communication, 1969) for quite some time as a
conglomeration of distinct pathophysiologic disorders culminating
in fasting hyperglycemia, diabetes mellitus has only recently
absorbed any particularly formal recognition of its incipient
complexity. In the otherwise unremarkable year of 1979, the
National Diabetes Data Group published its "Classification
and Diagnosis of Diabetes Mellitus and Other Categories of
Glucose Intolerance" (Diabetes 28(12):10039, 1979). That
landmark "consensus" divided overt diabetic territory
into two more homogeneous areas - Type 2 Diabetes (Non-Insulin
Dependent Diabetes Mellitus) and Type 1 Diabetes (Insulin
Dependent Diabetes Mellitus), and one heterogeneous gemisch
ignominiously labeled - "Other".
Type 1 Diabetes does appear to
represent the most homogeneous of these three categories, but
only in so far as its definition comprises absolute, isolated
ß-cell insulin secretory failure (Srikanta S, Ganda OP, Gleason
RE, Kaldany A, Garovoy MR, Milford EL, Carpenter CB, Soeldner JS,
and Eisenbarth GS, Ann Int Med 99:320, 1983). Most of the
patients in this category are young and manifest their disease on
the basis of a specific immunologic attack upon the pancreatic
ß-cell of the islets of Langerhans (Cudworth AG, Diabetologia
14:281, 1978).
The "Other" category
includes generalized islet failure, usually in the setting of
pancreatic exocrine deficiency and chronic pancreatitis of varied
etiologies. It includes as well, however, resistance to the
action of insulin caused by certain genetic disorders or by
excess concentrations of known counter-regulatory hormones, such
as epinephrine, cortisol, glucagon, and growth hormone.
Type 2 Diabetes accounts for
90-95% of the total number of patients with diabetes mellitus
(Harris MI, in National Diabetes Data Group, eds. Diabetes in
America:diabetes data compiled in 1984. Bethesda, Md.:National
Institutes of Health, Chapter VI, 1985: publication no. 85-1468).
It is certainly less
homogeneous than Type 1 Diabetes, although the majority (70%) of
patients who comprise Type 2 Diabetes manifest obesity as some
basis of their pathophysiology. The evidence for this is: (1) not
only are they obese at diagnosis (Everhart J, Knowler WC, and
Bennett PH, in National Diabetes Data Group, eds. Diabetes in
America: Chapter IV, loc.cit.), but (2) their disease dissipates
with weight loss significant to the order of 5.00 to 10.00 kg-2
of Body Mass Index (BMI) (Bistrian BR, Blackburn GL, Flatt JP,
Sizer J, Scrimshaw NS, and Sherman M, Diabetes 25:494, 1976\;
Genuth SL, Am. J. Clin. Nutr 32:2579, 1979\; Hughes TA, Gwynne
JT, Switzer BR, Herbst C, and White G, Am J Med 77:7, 1984\;
Hadden DR, Montgomery DAD, Skelley RJ, Tremble RR, Weaver SA,
Wilson FA, and Buchanan KD, Br med J 3:276, 1975\; Bogardus C,
Ravusin E, Robbins DC, Wolfe RR, Horton ES, and Sims EAH,
Diabetes 33:311, 1984) and (3) it is more difficult for obese
Type 2 Diabetes patients to lose weight than their obese
non-diabetic counterparts (Henry RR, Wiest-Kent TA, Scheaffer L,
Kolterman OG, and Olefsky JM, Diabetes 35:155,1986). Nutritional
status is therefore a key pathogenic component of a majority of
patients with Type 2 Diabetes.
Nevertheless, insufficient
insulin secretion to overcome hepatic [and peripheral] insulin
resistance seems to be the common pathophysiologic thread weaving
its way throughout disorders currently classified as Type 2
Diabetes (Haffner SM, Stern MP, Hazuda HP, Mitchell BD, and
Patterson JK, NEJM 319:1297, 1988). Since the genetic/concordance
data are much stronger for Type 2 Diabetes, than for Type 1
Diabetes (Newman B, Selby JV, King MC, Slemenda C, Fabsitz R, and
Friedman GD, Diabetologia 30:763, 1987\; Barnett AH, Eff C,
Leslie RD, and Pyke DA, Diabetologia 20:87, 1981), investigators
are feverously searching for the molecular basis of the defect
[or defects] in Type 2 Diabetes. Obesity itself, though, has also
been shown to have a very strong genetic predisposition (Bouchard
C, Tremblay A, Despres JP, Nadeau A, Lupien PJ, Theriault G,
Dussault J, Moorjani S, Pinault A, and Fournier G, NEJM 322:1477,
1990\; Stunkard AJ, Harris JR, Pedersen NL, and McClearn GE, NEJM
322:1483, 1990).
Given the above data relating
obesity to Type 2 Diabetes, perhaps a significant genetic link
which has been observed in nondiabetic offspring of Type 2
Diabetes patients might associate more closely with obesity
[which in turn correlates with hyperinsulinemia] than with
glycemia, per se (Haffner et al, Table 1, loc.cit). In any case,
what might make a particular obese individual predisposed to Type
2 Diabetes? In addition, is it possible that the route to Type 2
Diabetes could represent a normal physiologic escape from what
otherwise might develop into the pathophysiologic state of
"morbid obesity?" Are there any lines of evidence which
might lend support to such a hypothesis? Finally, or, perhaps
first of all, what is the nature of the relationships which exist
between insulin resistance, hyperinsulinemia, and obesity?
Epidemiology and
Cell Biology
For both the power and
direction of any light shed upon the world of Type 2 Diabetes to
date, we are certainly indebted to the National Diabetes Data
Group as a whole, and to such epidemiologists, particularly, as
Kelly West, Maureen Harris, Peter Bennett, Clifton Bogardus,
William C. Knowler, Stephen Lillioja, Michael Stern, Steven
Haffner, Paul Zimmett, Gary Dowse, Susan Serjeantson, Harry Keen,
Giancarlo Viberti, Per Bjorntrop, Jaako Tuomilehto, Ron Klein,
Larry Rand, Zoltan Skrabalo, Jean-Pierre Felber, Eveline
Eschwege, Maureen Harris, and Elizabeth Barrett-Connor, among
others.
Without the light of the above
epidemiologists, most of us would still be groping. Given the
genetic predisposition of Type 2 Diabetes, both prospective and
retrospective studies have focused, are focusing, and will likely
continue to focus upon the following populations: 1) Pima Indians
from the Gila River Valley 2) Nauruans from Polynesia and
Micronesia 3) Mexican Americans from San Antonio 4) Mauritians
from Mauritius 5) Maltese from Malta and Canada 6) Indians from
Fiji, Asia and North America (Cherokee,Choctaw) 7) Aborigines
from Australia 8) First degree relatives of patients with
documented Type 2 Diabetes.
The two common denominators
amongst all of the above groups has been (1) the statistically
significant highest incidences and risks of development of Type 2
Diabetes seen in any of the world populations and (2) the
presence of prandial hyperinsulinemia and/or insulin resistance
in the normal-glucose-tolerant first degree relatives of patients
with Type 2 Diabetes. Insulin-resistance is defined here as
decreased whole body glucose disposal under
euglycemic-hyperinsulinemic "clamped" conditions. The
increased incidence of Type 2 Diabetes in these populations with
a background of insulin resistance has been ascribed, not
altogether facetiously, by Koessler and others, to
"Coca-Colonization". Herein lies a good part of the
genesis and rationale of the "stuffed cell hypothesis."
Once the high incidence of Type
2 Diabetes in these groups was noted by the respective
epidemiologists, the genetic linkage was confirmed by documenting
an even higher association with Type 2 Diabetes in one or two
parents of [primarily those most obese] affected probands
(Knowler WC, Pettitt DJ, Savage PJ, and Bennett PH, Am J
Epidemiol 113:144, 1981). Obesity by itself is also considered a
very "strong risk factor for Type 2 Diabetes"
(Everhart, loc.cit.). Nevertheless, the single strongest
predictor of Type 2 Diabetes in both univariate and multivariate
analyses would appear to be postprandial glycemia or
"Impaired Glucose Tolerance - IGT" (Bennett PH, Knowler
WC, Pettitt DJ, Carraher MJ, and Vasquez B, In: Eschwege, E.,
ed., Advances in Diabetes Epidemiology. Amsterdam:Elsevier
Biomedical, 1982, p.65).
In a multivariate analysis of
Mexican-Americans, fasting glycemia in the presence of fasting
hyperinsulinemia and IGT was the best predictor of Type 2
Diabetes (Haffner SM, Stern MP, Mitchell BD, Hazuda HP, and
Patterson JK, Diabetes 39:283, 1990). Peter Bennett's study of
the Pimas reported eight years previously (Bennett, 1982,
loc.cit.) determined that those individuals who progressed from
IGT to Type 2 Diabetes had not only elevated fasting levels of
insulin and glucose, but significantly decreased postload insulin
concentrations as well. Unfortunately the Haffner group did not
measure postload insulin levels (loc.cit.,1990).
However Bennett's postprandial
insulin observations have found ample confirmation elsewhere
(Sicree, RA, Zimmett, P, King HOM, and Coventry S, Diabetes
36:179, 1987\; Charles MA, Fontbonne A, and Eschwege E,
Diabetologia 31 [Suppl1]:479A, 1988\; Kadowaki T, Miyake Y,
Hagura R, Akanuma H, Kuzuya N, Takaka F, and Kosaka K,
Diabetologia 26:44, 1984). This kind of data allowed Bennett to
postulate in 1982 (loc.cit.) that inadequate insulin secretion
might be a significant factor in the transition from IGT to Type
2 Diabetes! Nonetheless, in only one long-term (10 year) clinical
trial (Sweden/open/tolbutamide plus diet/Sartor, G, Scherten B,
Carlstrom S, Melander A, Nordan A, and Persson, G, Diabetes
29:41, 1980) did intervention into IGT seem to negatively
influence the development of Type 2 Diabetes, whereas two trials
(Bedford/blind/tolbutamide or diet/Keen, H, Jarrett RJ, and
McCartney P, Diabetologia 22:73, 1982\; Whitehall/
blind/phenformin/Jarrett RJ, Keen, H, McCartney P In:Eschwege, E.
loc.cit., p.95) showed no influence of intervention whatsoever.
[Tolbutamide enhances insulin release whereas phenformin improves
peripheral insulin action.] If obesity, IGT (especially with
fasting hyperinsulinemia and prandial hypoinsulinemia), and
parental history predict Type 2 Diabetes in these already high
at-risk groups, what are the factors which predict IGT and/or
obesity? At this level we are essentially looking at probands
with normal glucose tolerance who are not obese, yet who are at
the highest risk of developing obesity, IGT, and ultimately Type
2 Diabetes. Haffner et al (loc.cit.,1988) evaluated non-diabetic
Mexican-Americans and found elevated fasting and prandial
insulins, which associated in step-wise fashion with the number
of parents with Type 2 Diabetes. These authors state, though,
that "it is surprising that there is relatively little or no
variation in body-mass index or body-fat distribution across the
various family history categories (Table I)." In point of
fact, however, the data in their Table I demonstrate
statistically and clinically significantly increased mean BMI's
among probands with one or both Type 2 Diabetes parent[s] as
compared to those with no Type 2 Diabetes parents. In any event,
the main point this paper establishes is that in a non-diabetic
population at high risk for Type 2 Diabetes, there is significant
familial hyperinsulinemia already present. Keep in mind that this
population comprises those non-diabetics with both normal and
abnormal glucose tolerance, however.
It was Lillioja et al
(loc.cit., 1987) who first elegantly demonstrated a profound
degree of familial insulin resistance clustering among mostly
glucose tolerant (only 10% had IGT ) non-diabetic Pima siblings.
Measuring glucose disposal at two levels of hyperinsulinemia
(MMax @ 400 mUm-2min-1 and MSubmax @ 40 mUm-2min-1) with
euglycemia clamped, they showed that profound insulin resistance
(MMax) was highly familial and independent of obeseness, VO2 max,
age or sex. A lesser degree of insulin resistance (MSubmax) did
correlate as well with degree of obesity, age and sex.
Finally, fasting insulin levels
correlated with fasting glucose, as well as sibship, degree of
obesity, age, and sex. Thus, at least three different levels of
familial insulin resistance may be suggested by this data: (1)
genotypic and profound, (2) hormonal, phenotypic, and less
profound, and (3) deterioration secondary to consequences of the
aforementioned and least profound of all. Bogardus C, Lillioja S,
Nyomba BL, Zurlo F, Swinburn B, Esposito-Del Puente A, Knowler
WC, Ravussin E, Mott DM, and Bennett PH (Diabetes 38:1423, 1989)
have detected an apparent trimodal distribution of the Mmax
defect in the Pimas. This has led him to postulate an autosomal
co-dominant mode of transmission involving two alleles at a
single locus. This kind of genetic defect should result in
dose-dependent inhibition of glycogen synthase activity, if
Bogardus is correct on all counts.
Jose Caro and his group
described significantly reduced insulin- stimulatable glucose
transport in muscle biopsies taken from morbidly obese,
non-diabetic as well as diabetic patients in eastern South
Carolina about to have some surgical procedure (Dohm GL, Tapscott
EB, Pories WJ, Dabbs DJ, Flickinger EG, Meelheim D, Fushiki T,
Atkinson SM, Elton CW ,and Caro JF, J Clin Invest 82:486,1988) .
"Transport of 3-O-methylglucose and 2-deoxyglucose was
stimulated approximately twofold by insulin in muscle from normal
nonobese subjects and stimulation occurred in the normal
physiological range of insulin concentrations.
In contrast to insulin
stimulation of 3-O-methylglucose and 2-deoxyglucose transport in
muscle from normal, nonobese subjects, tissue from morbidly obese
subjects, with or without Type 2 Diabetes, were not responsive to
insulin. Maximal 3-O-methylglucose transport was lower in muscle
of obese than nonobese subjects. Morbidly obese patients, with or
without Type 2 Diabetes, have a severe state of insulin
resistance in glucose transport." This defect was also
documented in individuals with normal glucose tolerance. However,
a majority of these patients may not go on to develop Type 2
Diabetes and/or may not be related to patients who have Type 2
Diabetes. Is this then, perhaps, a manifestation of the Msubmax
defect which associates with a "corpulent
predisposition"? If so, is the linkage causal or
consequential? Which of these individuals, if any, also have an
Mmax defect? Finally, the crucial question: do the patients
either who do already or who will eventually manifest Type 2
Diabetes have the Mmax defect so elegantly demonstrated in the
Pimas?
The Pima Indian Study group at
NIDDK had previously demonstrated that physiological insulin
resistance in that patient population correlated anatomically
with a defect in muscle biopsy glycogen synthase activity.
"Muscle glycogen content and glycogen synthase activity were
measured in percutaneous muscle biopsy samples obtained from the
vastus lateralis muscle before and after the euglycemic clamp
procedure. The results showed that muscle glycogen synthase
activity at the end of the euglycemic clamp was well correlated
with insulin-mediated glucose storage rates at both low (r =
0.50, P less than 0.02) and high (r = 0.78, P less than 0.0001)
insulin concentrations\; and also correlated with M[submax] (r =
0.66, P less than 0.001) and M[max] (r = 0.76, P less than
0.0001). Similar correlations were observed between the change in
muscle glycogen synthase activity and glucose storage rates and
M[max and submax].
The change in muscle glycogen
synthase activity correlated with the change in muscle glycogen
content (r = 0.46, P less than 0.03) measured before and after
the insulin infusions" (Bogardus C, Lillioja S, Stone K, and
Mott D, J Clin Invest 73:1185, 1984). This group has subsequently
ascribed this defect to a decrease in glycogen synthase
phosphatase or possibly to some non-suppressiblity of cyclic AMP
dependent ("A-") [protein] kinase (Kida Y, Esposito_Del
Puente A, Bogardus C, and Mott DM, J Clin Invest 85:476, 1990\;
Okubo M, Bogardus C, Lillioja S, and Mott DM, J Clin Endocrinol
Metab 69:798, 1989) If this represents the source of primarily
the Mmax defect, how is it different from the Msubmax defect
which supposedly correlates with obesity.
Since this glycogen synthase
defect does still significantly correlate with Msubmax [but less
well so than with Mmax], there should be some correlation
additionally with obesity. Is this association with obesity
primary and, if so, is the association causal or consequential?
The data are certainly compatible with a schema in which Mmax
precedes and predisposes towards obesity which results in some
defect in glucose transport or Msubmax. The data are also
compatible with a schema in which the Law of Mass Action applies
to the glucose back-up associated with a glycogen synthase defect
which in turn decreases glucose transport which results in
glucose shunting to the adipocytes and, ultimately, obesity.
Bogardus et al (Diabetes, 1989,
loc.cit.) feels simply that the data "suggest that if there
is a gene for insulin resistance in the Pima, those who carry it
are more likely to become obese." Or rather might any
association with obesity be a secondary reflection of a defect
more proximally related to the insulin receptor? If the latter,
is it upstream of the glucose transport signal [as potentially
intimated by the data of Caro et al]? In this scenario, however,
one might expect a much tighter correlation between Mmax and
Msubmax, whether or not obesity were present\; one might expect
that tighter correlation, that is, unless: 1) obesity itself were
etiologic and made some underlying propensity towards impaired
glucose transport more manifest and 2) Mmax and Msubmax were
unevenly affected by the upstream defect Finally, are the two
defects completely independent and additive? In an attempt to
clarify this issue, Shulman GI, Rothman DL, Jue T, Stein P,
DeFronzo RA, and Shulman RG, (NEJM 322:223,1990) have reported a
study in patients with Type 2 Diabetes utilizing a technique as
yet unsubstantiated and not generally accepted, i.e., nuclear
magnetic resonance spectroscopy. This study, on the basis of many
assumptions, calculations, and uneven weighting of the data,
purports to show in vivo a defect in muscle glycogen synthesis
present in lean Type 2 Diabetes patients not found in normal
controls.
Perhaps more to the point,
Eriksson et al have performed a nice study utilizing the
hyperinsulinemic- euglycemic [and hyperglycemic] clamp in (1)
Type 2 Diabetes patients, (2) first degree non-diabetic relatives
of Type 2 Diabetes patients, and (3) healthy non-diabetic
controls (Eriksson J, Franssila-Kallunki, A, Ekstrand A,
Saloranta C, Widen S, Schalin C, and Groop L, NEJM 321:337,
1989). This time, however, relatives were divided into those with
and without normal glucose tolerance and compared with Type 2
Diabetes kin and non-diabetic controls.
Insulin infusion rates were
roughly equivalent to Msubmax in the Lillioja study. Total-body
glucose metabolism was significantly reduced in all three
metabolically related groupings compared to controls: Type 2
Diabetes: 3.20 0.19 mg kg-1 min-1 IGT: 3.81 0.29 mg kg-1 min-1
NGT: 4.38 0.34 mg kg-1 min-1 CONTROL: 6.64 0.55 mg kg-1 min-1
There were no differences between any of the groups with respect
to baseline glucose oxidation rate (as determined by indirect
calorimetry). Insulin-stimulated glucose oxidation rate only
differed significantly between Type 2 Diabetes patients (2.03
0.10 mg kg-1 min-1) and controls (2.91 0.23 mg kg-1 min-1).
Subtracting presumed
"oxidative glucose metabolism" from "total-body
glucose metabolism" supposedly yields "non-oxidative
glucose metabolism" which "almost completely accounted
for" the defect in total-body glucose metabolism [assuming
that the calorimetry data is reliable cf. Ferrannini E,
Metabolism 37:287, 1988]: Type 2 Diabetes: 1.18 0.17 mg kg-1
min-1 IGT: 1.40 0.22 mg kg-1 min-1 NGT: 1.81 0.27 mg kg-1 min-1
CONTROL: 3.76 0.55 mg kg-1 min-1 This non-oxidative glucose
metabolism defect persisted in normal glucose tolerant (NGT)
first degree relatives of Type 2 Diabetes despite a totally
normal first phase (0' -> 10') insulin secretion: Type 2
Diabetes: 7 7 U/ml 10 min IGT: 44 18 U/ml 10 min NGT: 110 13 U/ml
10 min CONTROL: 101 34 U/ml 10 min and an enhanced second phase
(10' -> 120') insulin secretion: Type 2 Diabetes: 50 19 U/ml
110 min IGT: 324 50 U/ml 110 min NGT: 465 40 U/ml 110 min
CONTROL: 325 93 U/ml 110 min.
Maximal stimulated insulin
secretion mirrored second phase responses for all categories as
well. First degree relatives who showed a defect in first phase
insulin response, however, appeared to manifest at least IGT in
these studies. Basal hepatic glucose outputs correlated
positively with fasting glucose values (r = 0.538, P_0.001) in
the subjects as a group, were significantly lower in all groups
than in Type 2 Diabetes, Type 2 Diabetes: 2.62 0.13 mg kg-1 min-1
IGT: 1.90 0.10 mg kg-1 min-1 NGT: 1.89 0.06 mg kg-1 min-1
CONTROL: 1.89 0.06 mg kg-1 min-1, and were suppressible by
Msubmax insulin infusion in all groups except Type 2 Diabetes
(N.B. negative values): Type 2 Diabetes: 0.31 0.09 mg kg-1 min-1
IGT: -0.27 0.19 mg kg-1 min-1 NGT: -0.72 0.29 mg kg-1 min-1
CONTROL: -1.83 0.50 mg kg-1 min-1
The
"Gluconeogenesis Suppression Fallacy"
It should be mentioned at this
juncture that "hepatic glucose output" as described in
most of the above studies, is a purely ersatz value which has
been calculated based upon the so-called "Steele
Equation". This equation in turn is predicated upon a few
assumptions, i.e., that there is (1) but a single pool (2) which
is some fraction (usually ) of the total extracellular fluid
space through which a given compound is (3) rapidly and (4)
evenly distributed, and that (5) all "specific"
radioactivity within that pool at any point in time represents
solely intact compound.
Given those caveats, the
equation derived by Robert Steele and his cohorts at Brookhaven
National Library and the New York University School of Medicine
(Amer J Physiol 187:15,1956), states that:- 1) Ra = F - Bt(
dit/dt) it and 2) Rd = Ra - dBt/dt where, Ra is the rate of
appearance of compound into the pool, Rd is the rate of
disappearance of the compound from the pool, F is the rate of
compound isotope infusion, Bt is the total amount of compound in
the pool at time "t", and it represents the specific
activity of the pool at time "t". Under most conditions
the rate of appearance of a compound should represent the rate of
infusion of the compound, Ri, plus the intrinsic production rate
of the compound, Rh, or: 3) Ra = Ri + Rh Rearranging, then: 4) Rh
= Ra - Ri
If the compound under
discussion is glucose, Rh should represent predominantly
"hepatic glucose output" [as the liver is the major
site of gluconeogenesis]. But how can an Rh be negative? That is
to say, under conditions of equilibrium, (and remembering that
appearance rates are independent from disappearance rates) how
can glucose appear [into whatever its putative distribution
space] at a rate slower than which it is infused? At equilibrium,
when Rh = 0 and dit/dt = 0. 5) Ra = F-Bt(dit/dt) = Rh + Ri it 6)
F/Ri = it 7) i = it [by definition] In order for Rh to be
negative in 4), Ra has to be UNDERestimated [i.e., Ra < Ri] by
either: a) if dit/dt <0, then EITHER its negativity must be
UNDERestimated or the pool size (Bt) UNDERestimated or b) if
dit/dt >0, then EITHER its positivity must be OVERestimated or
the pool size (Bt) OVERestimated Under conditions where Rh is
presumably equal to 0, a negative dit/dt must represent further
dilution of specific activity by a positive gluconeogenesis which
itself is either UNDERestimated or with a glucose pool size which
is UNDERestimated. Since Rh becomes even more negative with
higher rates of insulin infusion, either increasing insulin
concentrations increase the total glucose available to the
extracellular pool at any point in time (thereby leading to an
UNDERestimate of pool size) OR increasing insulin concentrations
magnify the ability of sampled slope estimates of glucose
specific activity to UNDERapproximate actual slopes at any point
in time. [If dit/dt were positive, then the specific activity of
the pool sample must be increasing with time at the moment of
sampling. "Delayed mixing" has been bandied about as
some sort of rationalization for this kind of result.
Nevertheless, this type of
explication fails to explain how Rh becomes even more negative
with higher rates of insulin infusion. If supposition
"b" is to be correct, then either the pool size for
glucose decreases as insulin increases, or the specific activity
of plasma glucose is enhanced by some non-glucose radioactive
label which has been stimulated by insulin, e.g., 14CO2, or
14C-[tri]glyceride, or 3H20, or 3H2CO3 or etc.] Some
investigators (e.g., Shulman et al, loc cit) who utilize the
glucose clamp technique base their assumptions that hepatic
glucose output is suppressed under conditions of hyperinsulinemia
and hyperglycemia ("Under these conditions - constant
hyperglycemia and hyperinsulinemia - hepatic glucose output is
completely suppressed8....") upon DiFronzo RA, Ferrannini E,
Hendler R, Wahren J, and Felig P, Proc Nat Acad Sci 75:5173,
1978. The superscript -8- here is just such a reference.
Nevertheless, a perusal of this
citation reveals absolutely no support for such an assumption.
Indeed, "splanchnic output" per se is reported under
basal conditions, but only net "splanchnic uptake" is
reported under conditions of hyperglycemia and hyperinsulinemia.
Thus, it is impossible to reach any conclusions about hepatic
glucose output under hyperglycemia and hyperinsulinemia from the
cited data provided by DeFronzo et al. However, this particular
reference does shed significant light upon hepatic
("splanchnic") glucose uptake under varied conditions
of insulinemia, glycemia, and route of glucose administration
(p.o. or i.v.).
Interestingly enough,
intravenously induced hyperglycemic clamping at 223mg/dl without
exogenous insulin [achieving comparable levels of endogenous
peripheral insulinemia (55U/ml) to that after a 100g oral GTT]
increased the total percentage of glucose taken up by the
splanchnic bed from ~5% [seen under euglycemia/hyperinsulinemia)
to ~14%. When this was repeated with exogenous hyperinsulinemic
clamping at the same level of hyperglycemia, the splanchnic
component decreased to ~9% of the total body glucose uptake. Now
when the glucose was administered orally [as well as
intravenously], net splanchnic uptake increased almost sixfold
from basal (DeFronzo et al, ibid.) and amounted to almost 60% of
the total body glucose uptake (Felig et al, Diabetes 24:468, 1975
and Diabetes 27:121, 1978). Yet, virtually all of the studies
purporting to show that the primary site of insulin resistance in
Type 2 Diabetes is in the periphery have been done under
conditions of solely intravenously administered glucose.
If hepatic uptake can amount to
at least 60% of an orally administered glucose load, and glucose
is almost always ingested orally, 1) what is the significance of
resistance demonstrated after solely intravenously administered
glucose? and 2) why have there not been more studies utilizing
clamping after orally as well as intravenously administered
glucose? Steele R, Bjerknes BA, Rathgeb I, and Altszuler N
(Diabetes 17:415, 1978) have shown that under normal physiologic
conditions during a 3 hour oral GTT in dogs, hepatic glucose
output is 54% of basal\;whereas total body glucose disposal is
163% of basal. This amounts to a net decrease of 0.29 g/kg/3hrs
in hepatic output (still positive at 0.34 g/kg/3hrs), and an
increase of 0.40 g/kg/3hrs in total body uptake (1.03g/kg).
Now, if DeFronzo et al are
correct in that the net splanchnic uptake during intravenously
induced hyperglycemia was ~14%, and Steele at al are correct that
under "supposedly" similar conditions of [peripheral]
insulinemia, hepatic glucose output was 0.34 g/kg/3hrs, and if
one can analogize somewhat from dog to man, then the hepatic
glucose uptake per se during those conditions in DeFronzo could
represent approximately 0.4842 g/kg/3hrs (14% of a total glucose
uptake of 1.03 mg/kg/3hrs added to a hepatic glucose output of
0.34 mg/kg/3hrs) or 47% of total body glucose uptake! Felig et al
(Diabetes 24:468, 1975\;Diabetes 27:121, 1978) show net
splanchnic uptakes of ~60% during 100 g oral glucose loads in
normal subjects. These would translate into hepatic output
components [during much more normal physiologic manipulation] of
approximately 93%! Bear in mind, however, that this figure is
based upon a rather sizeable glucose output under these
conditions as determined by Steele et al (loc. cit.).
Under conditions of exogenously
clamped hyperinsulinemia and hyperglycemia in DeFronzo (loc.
cit.), the net fractional splanchnic uptake dropped to ~9% [from
14%]. Even assuming (without the benefit of any experimental data
Shulman to the contrary not withstanding) that hepatic glucose
output is completely suppressed under these conditions [of
hyperglycemia and hyperinsulinemia], then hepatic glucose uptake
would be equal to the net splanchnic uptake, and decrease to
approximately 9% (from 47%) of total body glucose uptake.
Thus, increasing insulin to
maximum levels [roughly four times those seen endogenously during
an oral glucose load] could appear to increase predominantly
peripheral (i.e., non-hepatic) glucose uptake. Might one not
wonder as well whether any "peripheral insulin
resistance" determined under these somewhat artificial
conditions were conceivably a bit over-represented - and
"hepatic insulin resistance" likewise somewhat
under-represented?
The net upshot of all of this
is that:- 1) Gluconeogenesis may be [significantly]
underestimated in some studies (v.s.) due to either inaccurate
pool size estimates or inaccurate accounting of plasma isotope
activity 2) Hepatic glucose uptake may be [significantly]
under-represented in some studies (v.s.) due to non-physiologic
excesses of insulin [glucose] and failure to accurately account
for gluconeogenesis under all conditions 3) Peripheral glucose
uptake may be [significantly] over- represented in some studies
(v.s.) due to non-physiologic excesses of insulin [glucose] and
failure to accurately account for gluconeogenesis under all
conditions
To further obfuscate the
picture DeFronzo RA, Ferrannini E, and Simonson DC have
apparently calculated from their data (Metabolism 38:387, 1989)
that basal hepatic glucose output is not significantly different
in "mild" diabetics from non-diabetic controls. This is
in distinct contrast to conclusions drawn by Efendic S, Karlander
S, and Vranic M (J Clin Invest 81:1953, 1988) who showed markedly
increased basal hepatic glucose outputs and markedly enhanced
intrahepatic glucose cycling in their patients with mild
diabetes. Moreover, young non-diabetic Australian aborigines
apparently have significantly greater basal and significantly
less suppressible hepatic glucose outputs in response to oral
glucose loading than controls (Proietto J, Nankerois K,
Traicenedes K, Rosella J, and O'Dea K, Diabetes Res Clin Pract 5
[Suppl.I]:263, 1988).
Finally, using glucose labeled
with tritium, and assaying for glucose specific activity
utilizing the perchloric acid precipitation technique of Best et
al (Diabetes 30:847, 1981), Kwame Osei (Diabetes 39:597, 1990)
has shown that hepatic glucose outputs were significantly greater
(25%) in nondiabetic first-degree relatives than in controls
(which correlated in statistically significant fashion with
fasting plasma glucose and fasting plasma c-peptide). Remember,
however, that there did not appear to be any differences in basal
glucose outputs from Erikkson et al (loc cit) in either normal-
(N = 13) or impaired- glucose tolerant (N = 13) first-degree
relatives from controls (N = 14). Even though more statistical
power (N = 27 proband vs. N = 16 control) was available to detect
this difference in Osei (loc cit\; p _ 0.05), further
corroboration is certainly desireable.
The data of Niewohner CB and
Nuttall FQ (Diabetes 37:1559, 1988) in rats "suggest rapid
bidirectional glucose flux" into and out of the liver
following an oral glucose load, "but with active transport
maintaining a modestly higher intracellular glucose
concentration" [than the extracellular milieu].
The data supplied by Reaven's
group at Stanford entitled "Effect of Central Obesity on
Regulation of Carbohydrate Metabolism in Obese Patients with
Varying Degrees of Glucose Tolerance (Golay A, Chen N, Chen YDI
Hollenbeck C, and Reaven GM, J Clin Endocrinol Metab 71:1299,
1990) may shed additional support for this concept. In Figure 2
of that paper, there were 6 patients with glucose disposals
(Rd's) in the range of 1.2 to 1.75 mmoles m-2 min-1: 3 patients
with NGT and 3 patients with IGT. The means of both groups were
roughly equal [if anything the IGT patients trended a bit
higher]. Assuming the group diagnoses based on oral glucose
tolerance characterizations to be correct in those 6 patients, if
the glucose disposals were equivalent, then those of these
patients with IGT must have manifested increased glucose outputs
as the primary basis of their abnormal glucose tolerance.
Widening the Rd window from 1.0
to 2.0 mmoles m-2 min-1 reveals 11 patients, 5 with NGT, 5 with
IGT, and even 1 with Type 2 Diabetes. Roughly equivalent (with
NGT patients this time trending higher than IGT), the means of
the Rd's again imply that the differences in glucose tolerance
must be accounted for primarily by differences in glucose outputs
- compatible with current concepts of Type 2 Diabetes, but a
rather new observation in IGT. The three other NGT patients
included in this figure had Rd's greater than 2.5 mmoles m-2
min-1 (four NGT patients were not plotted for some reason). Three
other IGT patients, and 4 other Type 2 Diabetes patients (six not
plotted) all had Rd's less than 1.0 mmole m-2 min-1. The means
for the three groups as a whole with standard deviations,
however, appeared to be as follows: Dx Rd NGT: 2.09 0.27 mmoles
m-2 min-1 IGT: 1.14 0.12 mmoles m-2 min-1* Type 2 Diabetes: 0.80
0.09 mmoles m-2 min-1** *p < 0.01 compared to NGT **p <
0.001 compared to NGT.
Nevertheless, even assuming
that these group means are accurate, the take home message here
is that a significant percentage (_62.5%) of obese patients with
IGT may have increased [hepatic or renal] glucose outputs as the
primary basis for the impairment in glucose tolerance, i.e.,
assuming insulin responses are comparable. Even though individual
points for insulin responses and Rd were not shown in this
report, the total mean daily integrated insulin response for the
IGT group as a whole (3967 394 pmol L-1 h-1) was almost twice as
high as that of the NGT group (2418 243 pmol L-1 h-1).
Although certainly not
conclusive, it is therefore quite likely that glucose output was
the major contributor to the impaired glucose output seen in this
obese population. [It might also be noteworthy here to point out
that the above authors also appeared to be somewhat reluctant to
accept the notion that hyperglycemic/hyperinsulinemic conditions
necessarily shut-down glucose output:- "If endogenous
glucose production is totally suppressed during the insulin clamp
study," [author's note: no references cited] the infusion
rate of glucose needed to maintain the plasma glucose at 14mmol/L
provides an estimate of the glucose disposal rate."
Therefore, these authors used glucose with tritium labeled in the
C3 position (in conjunction with the Steele equation) to estimate
glucose disposal rate or Rd.]
I believe the notion that: the
percent of hepatic glucose output suppression EXCEEDS the percent
of maximal peripheral glucose uptake at any given and equal
concentration of insulin [and glucose] in portal and peripheral
capillary beds should be termed the "gluconeogenesis
suppression fallacy." In the first place, I know of no good
evidence to support such a notion. In the second place, the liver
usually sees much higher post-prandial concentrations of insulin
than does the periphery [due to the physiology of insulin release
directly into the portal circulation.] In the third place and in
a purely teleological sense, it would seem prudent from the
perspective of the body's major glucose utilizer, i.e., the CNS,
for gluconeogenesis [glucose-6-phosphatase?] to be somewhat
resistant to the insulin message.
Therefore, it would seem much
more likely and physiologically appropriate for: the percent of
maximal peripheral glucose uptake to EXCEED the percent of
hepatic glucose output suppression at any given and equal
concentration of insulin [and glucose] in peripheral and portal
capillary beds. Epidemiological data (Keen, H, Jarrett RJ, and
McCartney P, Diabetologia 22:73, 1982\; Jarrett RJ, Keen, H,
McCartney P In:Eschwege, E. loc.cit., p.95.) support the notion
that most Type 2 Diabetes (51%) and IGT patients [but not the
Pimas] have a very high prevalence of Type IVa or IVb
hyperlipoproteinemia (20% have IIa) which, in turn, correlates
very strongly with macrovasculopathy and coronary mortality [but,
again, not in the Pimas].
Reaven and others have
attributed this phenomenon to increased hepatic synthesis of
triglycerides due to enhanced insulinemia. However,
hyperinsulinemia may increase VLDL clearance as well by
increasing [non-myocyte] lipoprotein lipase (LPL) synthesis
(Garfinkle AS, Nilsson-Ehle P, Schotz MC, Biochim Biophys Acta
424:265, 1976). Nevertheless, insulin may also suppress muscle
LPL synthesis thereby shunting VLDL uptake to adipocytes (Dobs
AS, Prev Cardiol Rpts 4:1, 1991)
The obvious question here
relates to the possibility of differential insulin sensitivities
as well as differential hepatic effects of insulin upon fatty
acid, amino acid, and glucose metabolism, i.e., if hepatic
insulin sensitivity were greater than peripheral sensitivity
(relative to ambient insulin concentrations), then would not one
expect hyperlipemia in the face of IGT? More especially would one
not expect hyperlipemia and IGT during fasting conditions of
basal [hyper]insulinemia? Differential hepatic effects of insulin
upon glucose and lipids might also explain how hyperlipemia could
associate with fasting hyperglycemia and Type 2 Diabetes (once
fasting insulin levels drop sufficiently). On the other hand, if
both hepatic and peripheral sensitivities were decreased more or
less equally (relative to ambient insulin concentrations), then
might one not expect IGT (or Type 2 Diabetes depending on basal
levels of insulinemia) without hyperlipemia [as in the Pimas, or
treated Type 1 Diabetes, or later stage Type 2 Diabetes
patients]? It is also possible, however, that the Pimas could
have a glycogen synthetase defect in the periphery and,
potentially, liver, as well as having either a lipogenic defect
alone in the liver or lack of insulin suppressibility of LPL
activity in muscle. In summary, then, Caveat Emptor! Although the
preponderance of current dogma [based on IV infusions of both
glucose and insulin] might favor a primary defect in insulin
resistance peripherally at the level of the myocyte and glycogen
synthesis and/or glucose transport, there does not appear to be
any convincing data that a primary defect might not occur
centrally at the level of the hepatocyte with, perhaps, increased
glucose cycling. A primary defect in both muscle and liver
glycogen synthase activity [or in enhanced hepatic
gluconeogenesis] could shunt more of the glucose which is then
taken up by adipocytes [and hepatocytes] in the direction of
enhanced lipogenesis. "Cell
Stuffing","Left-Shifting" and
"Uncoupling" - A Basis for Ongoing Scientific
Investigations The concept that an "ingestion
threshold" might exist at the cellular level is herein only
somewhat facetiously referred to as "cell stuffing".
There is no little experimental basis for this concept which,
indeed, may turn out to be rather key in glucose homeostasis,
particularly as relating to the physiology of glucose-dependent
insulin secretion by the ß-cell. Increases in intra(ß-)cellular
glucose concentrations are reflected by glycolytic and oxidative
phosphorylations of ADP to ATP. Higher levels of ATP in turn
inhibit potassium channels resulting in depolarization of the
plasma membrane. When this occurs, calcium channels open and
calcium influx (Arkhammar P\; Nilsson T\; Rorsman P\; Berggren
PO, Inhibition of ATP-regulated K+ channels precedes
depolarization-induced increase in cytoplasmic free Ca2+
concentration in pancreatic beta-cells, J Biol Chem 262 (12):
p5448-54, 1987) is associated with [?activation of another
G-protein and] phospholipase-C activation, conversion of
phosphatidylinositol 4,5-bisphosphate (PIP2) into inositol
1,4,5-trisphosphate (InsP3) and sn-1,2-diacylglycerol (DAG),
activation of protein kinase C, threonine and serine
phoshorylation, and a cascade ultimately resulting in insulin
release and synthesis. It turns out that in the presence of
protracted and relatively-mild hyperglycemia, this pathway
becomes less sensitive to glucose (Sako Y\; Grill VE Coupling of
beta-cell desensitization by hyperglycemia to excessive
stimulation and circulating insulin in glucose-infused rats.
Diabetes 39 (12): p1580-3, 1990). It further turns out that upon
restoration of euglycemia, [and in the absence of genetically
susceptible irreversibility and diabetes (Korsgren O\; Jansson
L\; Sandler S\; Andersson A,Hyperglycemia-induced B cell
toxicity. The fate of pancreatic islets transplanted into
diabetic mice is dependent on their genetic background,J Clin
Invest 86 (6): p2161-8, 1990)] this pathway can restore itself
virtually completely to normal (Leahy JL Bonner-Weir S, and Weir
GC, J Clin Invest 81:1407, 1988).
Referred to recently as
"glucose toxicity", the limiting step in this process
may not unlikely be an inverse function of the potential change
from baseline to Vmax of ATP generation via glucose catabolism,
i.e., the higher the baseline ATP generation, the lower the
secretory response to a given glucose challenge. This same
desensitization as a manifestation of the Pasteur effect could
further feedback inhibit glucose uptake [with partially
corrective repercussions]. "Glucose toxicity"
apparently does not result in a malicious cycle of increasing
decompensation as one might, perhaps, have otherwise expected
(Weir GC, Joslin Symposium on Type 2 Diabetes, Boston, MA,
October 1, 1990)
Now I must unfortunately
disappoint Sir William of Occam by introducing a second
hypothesis, i.e., "left-shifting of the glucose-ATP/insulin
dose response curves". Usually, when one deals with the
physiopathology of drug, hormone, ligand failures, one speaks in
terms of "right-shifting of the dose response curves",
i.e., it takes more of some agonist to induce a given effect.
More correctly, the concentration of agonist resulting in
half-maximal activity [Km] turns out to be higher, and, thus
"right-shifted". Furthermore, pharmacologic
intervention usually involves some attempt to
"left-shift" the dose-response curve for some agonist
which of necessity requires more potency in order to achieve its
desired biochemical or physiologic effect.
Coming back to glucose, ATP,
and the ß-cell, then, several questions arise: 1) Why do many
patients who have had gestational diabetes manifest decreased 1st
phase insulin responses when not pregnant? (Porte,D, Seminar at
Molecular Diagnostics, West Haven, September 25, 1990) 2) Why are
normal, i.e., non-diabetic pregnancies associated with such a
marked lowering of basal glycemia in the face of quite markedly
increased insulin resistance and comparably increased insulin
requirements? 3) Why does reactive hypoglycemia seem to be
associated with impaired glucose tolerance and early diabetes
(Khurana et al, Postgrad. Med. 53: 118,1973), and more especially
in obese, first degree relatives of patients with Type 2 Diabetes
? (Ensick JW and Williams RH, p.858 in Williams RH, ed., Textbook
of Endocrinology, W.B. Saunders, Philadelphia, 1981 citing the
work of LeFebvre et al in 1976) 4) Might there not be some common
explanation for the above observations? 5) Assume, (for the sake
of argument), that agonists such as (a) estrogen progesterone or
(b) some as yet unidentified compound associated with obesity
could induce a left-shifting of the glucose-ATP dose response
curve. How might the natural history of subsequent events
develop? | <= intra ß-cell [ATP] | x x | x | x | x |
<--"threshold" level--> | x | x | x | x | x | |
+------------------+----+---+----------------+------------- 7.5mM
15mM ambient extra ß-cell [glucose] x <=>
"agonist" <=> "normal" (+ <=>
glucose concentration at "threshold") Figure 1 The
early natural history of this kind of defect should reveal lower
basal levels of glycemia which border or even infringe upon the
limits of hypoglycemia as seen in normal gestation. A strong
predisposition for weight-gain should result. This could further
exacerbate insulin resistance. If meals were eaten so frequently
or insulin resistance increased to such an extent that resultant
elevated ambient glucose concentrations reflected
intra[ß]cellular ATP concentrations increasing above the dotted
line (Fig.1), then the decreasing slope of the dose-response
curve could represent the early limits of 1st phase insulin
secretory failure. What is intriguing is that, as with gestation,
the threshold glucose concentration for this secretory failure
could be extremely low (on the order of 6mM/L or 102 mg/dL).
Thus, and in summary, a
"left-shifting" of the glucose-ATP dose response curve
should result in:
1) Hypoglycemia which is: a) Early and basal during
gestational periods b) Reactive in the presence of obesity or
insulin resistance
2) 1st phase insulin resistance [if ambient glucose concentrations
increase above some (minimal) threshold] which is: a) Completely
reversible if the curve can be "right-shifted" and the
ambient glucose concentration lowered below the threshold, e.g.,
by reducing insulin resistance b) "AT-RISK" if the
curve cannot be "right-shifted" and partially
reversible if the ambient glucose concentration can be lowered
below the threshold.
This is a situation like that
in previously gestational diabetics in which: (i) Fasting glucose
levels could still be well within normal limits (ii) High insulin
levels would imply a significant degree of insulin resistance
(iii) Persistent elevations in ambient glucose concentrations
could possibly induce further "left-shifting" of the
glucose-ATP dose-response curve and foster "uncoupling"
("right-shifting" with decreasing Vmax) of the
ATP-insulin dose-response curve (Fig.2) (iv) Low insulin levels
would particularly imply uncoupling of ß-cell ATP concentrations
from ultimate insulin release]. | <= insulin release | o o | o
| o | o | + + | o + | o | o | o + | o +
+----------------v--v-------------------------------------- Km's
ß-cell [ATP] o <=> "normal" ß-cell (function
and mass) + <=> "abnormal" ß-cell (function or
mass) AND prolonged above-threshold glycemia Figure 2 In order
for there to be ß-secretory failure according to this model,
then, two conditions would have to apply: 1) Left-shifting of the
glucose-ATP dose-response curve (Fig. 1) 2) Uncoupling
(right-shifting with increased Km and decreased Vmax) of the
ATP-insulin dose response curve (Fig. 2) Thus glucose-insulin
dose responses [I(g)] can be considered to be a function of both
the glucose-ATP [A(g)] and ATP-insulin [I(A)] dose-responses,
i.e., I(g) = A(g) + I(A) (Fig.3) | <= intra ß-cell [ATP] | o
o | o | | o | + + | o + | | o + | o | o + Km | |
+---------------------v-+--------------------+------------- 7.5mM
15mM ambient extra ß-cell [glucose] o <=> normal ß-cell
(function and mass) + <=> abnormal ß-cell (function or
mass) AND prolonged above-threshold glycemia Figure 3
If this model turns out to be
correct, we should then devote our efforts to designing
therapeutic modalities which "right-shift" and thereby
raise the Km of glucose for ß-cell ATP generation. Metformin is
one agent which could be a prototype or lead-structure. The net
effect of such a strategy would be to increase reserve ATP
generating capacity at any given effector concentration. This
strategy should be combined with a search for an agent that
recouples insulin secretion to ATP (e.g., sulfonylureas?) Does
this concept that an "ingestion threshold" might exist
at the cellular level apply to the periphery as well as well as
it does to the ß-cell?
Prolonged hyperglycemia in
adipocytes does seem to result in fairly substantial insulin
resistance. Moreover, this significant degree of insulin
resistance appears to be reversible with restoration of
euglycemia (Traxinger RR and Marshall S, J Biol Chem 264:8156,
1989). In addition, this effect is significantly aggravated by
glutamine in dose-dependent fashion (Traxinger RR and Marshall S,
J Biol Chem 264:20910, 1989). Fasting and weight loss (which
effect cellular "unloading" or "unstuffing")
both improve insulin secretory responses AND/OR peripheral
sensitivity to insulin in obese Type 2 Diabetes (Savage PS,
Bennion LJ, Flock EV, Nagulesparan M, Mott D, Roth J, Unger RH,
and Bennett PH) J Clin Endocrinol Metab 48:999, 1979\; Hidaka H,
Nagulesparan M, Klimes I, Clark R, Sasaki H, Aronoff SL, Vasquez
B, Rubenstein AH, and Unger RH, J Clin Endocrinol Metab 54:217,
1982\; Henry RR and Olefsky JM, Diabetes 35:990, 1986).
The hypothesis here is that -
cellular stuffing of the ß-cell on the one hand, as well as of
the myocyte, adipocyte, and hepatocyte on the other hand - could
ultimately result in (1) decreased insulin secretory responses to
glucose and (2) decreased peripheral glucose responses to
insulin, respectively [and by potentially similar
pathophysiologic mechanisms.] Indeed, in rats with already
abnormal ß-cell function (2 to streptozotocin), high fat feeding
alone without carbohydrate or protein [hepatocellular and
adipocyte stuffing] resulted in significant fasting hyperglycemia
(Pascoe WS and Storlien LH, Diabetes 39:226, 1990).
A cartoon speculating about the
possible natural history of the evolution from insulin resistance
to Type 2 Diabetes is presented as Figure 2. Vertically there are
three columns: 1) The left-most column portrays events putatively
transpiring in the insulin-responsive cellular compartment
comprising predominantly hepatocyte, adipocyte, and myocyte. 2)
The central column represents the physiologic state at each level
of development. 3) The right-most column shows events as they are
theorized to evolve in the ß-cell at any given stage.
Horizontally 5 stages are depicted: 1) The top level represents
normal glucose tolerance, insulin sensitivity, and insulin levels
and responses. 2) The second level reveals the state of insulin
resistance. 3) The third level describes the phase of impaired
glucose tolerance (IGT). 4) The fourth level shows the stage of
overt Type 2 Diabetes. 5) The final level attempts to portray
metabolic deterioration in Type 2 Diabetes. The central theses
implied herein are that: 1) Increased ingestion of carbohydrate
and protein ultimately lead to increased presentation of glucose
and glutamine to insulin-responsive cells. 2) These
insulin-responsive cells have an "ingestion-threshold"
which in turn is dependent on: a) Intracellular concentrations of
ATP b) Intracellular ratios of creatine-phosphate to ATP ratios
[myocytes] c) Total intracellular glycogen [myocytes and
adipocytes] d) Total intracellular triglyceride [adipocytes] e)
Total cell size [adipocytes] f) ?Intracellular concentrations of
cyclic adenosine monophosphate (cAMP)
3) Once these thresholds have
been reached and/or exceeded there is a right-shifting of the
insulin-net glucose utilization dose-response curve
("insulin resistance").
4) Increased ingestion of
carbohydrate also results in increased presentation of glucose to
insulin-secretory (ß-) cells.
5) Sustained elevated ambient
glucose concentrations (as well as female sex hormones and some
obesogenic factors) result in increased ß-cell basal
concentrations of ATP as well as a left-shifting of the
glucose-ATP dose-response curve.
6) Sustained left-shifting of
the glucose-ATP dose-response curves put a patient at increased
risk for 1st phase insulin secretory defects.
7) Sustained increases in
ß-cell basal ATP concentrations above a threshold level may
cause "fatigue" or uncoupling of the ATP-insulin
responses. The "apparent" Km should then increase with
right-shifting and Vmax depression of the ATP-insulin
dose-response curve. These changes ought inevitably to result in
not only diminished 1st phase insulin secretory responses, but
also in eventual ß-cell failure as well.
8) Hypersensitivity of this
uncoupling phenomenon might very easily represent the primary
ß-cell defect which is ultimately expressed as Type 2 Diabetes.
Insulin-Resistance
The phase of insulin resistance
may be due to any number of factors, either acting singly, or in
concert: 1) Some myocyte defect in the glycogen synthase pathway
(as defined in Pimas by the NIDDK group in Arizona) 2) Some
myocyte defect in glucose transport (as defined by the Caro group
at East Carolina State) 3) Some adipocyte defect in glucose
utilization perhaps secondary to glutamine suppression (as
described by Marshall's group in Memphis) 4) Some possible
prandial defect in suppression of hepatic glucose output (? by
glutamine or lactate) 5) Presence of "ingestion
threshold" conditions (v.s.) 6) Excess insulinemia with
normal to low glucose levels due to compensated left-shifting of
the glucose-ATP/ATP-insulin dose- response curves The
manifestations of such defects would be: 1) Decreased glycogen
synthesis by myocytes 2) Increased lipogenesis and VLDL clearance
by adipocytes and hepatocytes 3) Increased post-prandial
insulinemia 4) Prandial and fasting euglycemia
Impaired Glucose
Tolerance (IGT)
The development of impaired
glucose tolerance from insulin resistance is probably due to: 1)
Progression of insulin resistance 2) A 1st phase insulin
secretory defect possibly resulting from increased ß-cell basal
ATP concentrations above a threshold level and uncoupling of
insulin secretion from ATP generation The manifestation of these
progressive defects would be: 1) Decreasing lipogenesis and VLDL
clearance by adipocytes 2) Increasing lipolysis and free fatty
acid (FFA) production and release by adipocytes 3) Decreasing
hepatic glycogen synthesis 4) Increasing hepatic
re-esterification of FFA to triglyceride with enhanced VLDL
production rates 5) Decreasing post-prandial insulinemia 6)
Elevated fasting insulinemia 7) Prandial hyperglycemia with
fasting euglycemia
Non-Insulin-Dependent
Diabetes Mellitus (Type 2 Diabetes)
The evolution of Type 2
Diabetes from IGT may initially be reversibly due to consequences
of glucose toxicity by: 1) Progression of factors in Insulin
Resistance and IGT 2) Approaching the Vmax of the
glucose-ATP/ATP-insulin dose-response curves in the ß-cell 3)
Further left-shifting of the glucose-ATP and uncoupling of the
ATP-insulin dose-response curves in the ß-cell 4) Progressively
enhanced hepatic resistance to the insulin message possibly due
to a combination of: a) Increasing free fatty acid levels b)
Increasing lactate levels c) Progressive exceeding of
"ingestion threshold" limits 5) Fasting insulin levels
insufficient to overcome hepatic resistance The manifestations of
Type 2 Diabetes are: 1) Greatly increasing lipolysis and fasting
FFA production by adipocytes 2) Increasing hepatic fasting and
prandial gluconeogenesis 3) Decreasing fasting insulinemia 4)
Prandial and fasting hyperglycemia
Metabolic
Decompensation
Severe metabolic decompensation
may be progressively due to the unopposed effects of
"cell-stuffing" and "glucose toxicity" upon
both insulin target organs and insulin producing cells. It is
manifested by increasing fasting and prandial glucose levels in
the face of decreasing fasting and prandial insulin levels.
Amylin - an insulin-antagonistic peptide co-secreted
physiologically with insulin in a molar ratio of approximately
1:100 - may play a later role in a progressive, irreversible
ß-cell failure seen in the presence of amyloid deposition and
ultimate ß-cell necrosis.
Reload The National Diabetes Center
Frames
Home Page