Frederick M. (Frederick Madison) Allen.

Total dietary regulation in the treatment of diabetes online

. (page 65 of 76)
Online LibraryFrederick M. (Frederick Madison) AllenTotal dietary regulation in the treatment of diabetes → online text (page 65 of 76)
Font size
QR-code for this ebook


tolerance. The finding of such "atrophy" after widely different
dietetic management, i. e. in flagrant diabetes with no food restric-
tion whatever, and after the traditional protein-fat diet as in
Weichselbaum and Stangl's cases, and also in cases kept symptom-
free by prolonged rigid imdernutrition, furnishes partial proof, first,
that it is not of dietary origin, and second, that it is permanent and
irremediable under any form of treatment now known.

3. Cases in which the visible abnormalities in both islands and
acini seem too slight for a rational explanation of the diabetes. Some
of these have given rise to the behef that diabetes may exist with a
histologically normal pancreas. Some examples have been cases of
very rapid course in young persons, formerly supposed to represent
the extreme of severity, and therefore responsible for considerable
confusion in the past. Autopsies of such cases in their incipiency are
rare and therefore will be specially valuable whenever obtainable.
From some examples of early death in coma, however, it is possible
to infer that diabetes may begin with a pancreas showing only slight
fibrous or round cell invasion and containing an apparent abundance
of normal looking islands. One difficulty in forming quantitative
judgments is that the amount of pancreatic or island tissue necessary
to prevent diabetes in man is not known. Comparison with animals
is unreKable, since these vary widely and irregularly, from the cat at
one extreme, which becomes diabetic with about a fifth of the pan-
creas remaining, to the pig, in which a tiny fragment of pancreas
prevents both diabetes and cachexia. The monkey approximately
resembles the dog in requiring removal of seven-eighths to nine-tenths
of the pancreas for diabetes. Man is probably to be reckoned among
the most highly susceptible species, but the few reports of resection
of the human pancreas do not suffice for decision, partly because of



618 CHAPTER vin

inexact estimation of the portion removed or left, and partly because
the pancreas in such cases is diseased. Careful surgical, medical, and
pathologic studies of such patients in the future wUl possess obvious
importance. The best information at present seems to be afforded
by the case of Ghon and Roman," of a boy dying at the age of 14, with
idiocy, status lymphaticus, and other abnormalities, including a
congenital pancreatic deficiency, apparently the result of failure of
development of the dorsal pancreatic anlage, so that only the portion
formed by the ventral anlage was present. This roughly rectangular
plate of tissue adjoining the duodenum measured in its greatest
dimensions 7.4 cm. in length, 4.5 cm. in width, and 2 cm. in thickness.

Microscopically, "the changes consisted on the one hand in a focally occurring
increase of the interstitial tissue, in which here and there also small roimd cell
infiltrations were observable, and on the other hand in an everywhere observable
injury of the Langerhans cell islands. In comparison with the pancreas of a boy
of the same age dead of scarlet fever, studied for a control, the Langerhans cell
groups of our case showed not only increase of the intrainsular connective tissue,
but also the change which Weichselbaum has described as hydropic degeneration.
The hydropic degeneration change was mostly demonstrated only in its incipient
stages, but we saw also many islands, which following Weichselbaum we could
designate as atrophic."

The boy was physically well developed, and the existing diabetes
therefore probably of recent origin. The authors are doubtless jus-
tified in attributing its onset to the sUght pancreatitis rather than
to the simple metabolic strain of growth and puberty. There is a
further possible question whether the abnormalities in the body else-
where had any influence, and even whether the inherent functional
capacity of the islands was normal. On the whole, however, it ap-
pears that this considerable piece of pancreas was barely suflScient for
metaboHsm, and that diabetes ensued when it was only slightly in-
jured by inflammation. This interpretation agrees with other scanty
evidence that man is rather highly susceptible to diabetes, and that
the "margin of safety" in the human pancreas is not so wide as in
most experimental animals.

In addition to anatomic destruction, the existence of a functional
incapacity of the island cells must be assumed, and is demonstrated,
even without excessive labor of counting and measuring islands, by



ETIOLOGY ANT) PATHOLOGY 619

the following five facts: (a) Autopsy comparisons show that diabetes
often exists in cases where island tissue is widely and unmistakably
more abundant than in other cases without diabetes. (6) The ana-
tomic deficiency of islands is generally demonstrated only at death
from diabetes, after a course of months or years, during which pan-
creatitis or hydropic degeneration or both have been in progress;
and it must be assumed therefore that islands were more plentiful
at the beginning of the diabetes, (c) As illustrated by certain cases in
the present series and others,^' the most intense, even "total" dia-
betes may exist, as demonstrated by the dextrose-nitrogen ratio and
the respiratory quotient, and yet fasting may bring the condition
under control and a tolerance amounting sometimes to hundreds of
grams of carbohydrate may rapidly be recovered. Such patients
formerly met early death in coma, supposedly because of the hopeless
severity of their disease, and perplexity concerning the pathology
was created when they were found sometimes to possess a consider-
able abundance of apparently normal islands. It is now evident
that the recovery of assimilation under fasting would be impossible
if islands were not thus present; and these findings, formerly a
ground of argument against the insular hypothesis, actually serve to
support it. In the preliminary publications, it was deduced from this
fact alone that the deficiency of the islands must be at least partly
functional, (d) When some cells of the diabetic islands show hy-
dropic degeneration, a functional insufficiency must be assumed also
in the cells which appear normal. The accepted interpretation of the
hydropic change in both man and animals implies that the cells are
overtaxed functionally while they still appear normal, and the visible
vacuolation follows, (e) Though it is possible to find animal speci-
mens in which the exhaustion is universal, this condition has never
been observed in man. Even when the diabetes is not only maximal
by the usual tests, but furthermore so severe that it cannot be con-
trolled by fasting, the islands contain not only exhausted cells but also
others which appear structurally normal. The latter sort are fre-
quently so mmierous that a functional disabiHty must be admitted.

^ AUen, F. M., and Du Bois, E. F., Arch. Int. Med., 1916, xvu, 1010-1059. •



620 CHAPTER VIII

B. Changes Due to Diabetes.

The years 1900 and 1901 are epoch making in the microscopic
pathology of diabetes, for in them appeared not only Opie's study of
changes causing the disease, but also the first description by Weich-
selbaum and StangP^ of the hydropic degeneration, which later has
proved to be a result of the diabetes. Those authors described the
process as a vacuolation and liquefaction.

"There appear in the protoplasm of single or several epithelial cells of the
islands sniall vacuoles, or, more objectively expressed, small, round spots, within
which the protoplasm, which otherwise ordinarily shows a very close and fine
granulation and stains lightly but distinctly with eosin, appears entirely trans-
parent and colorless. These so called vacuoles then become confluent, and in the
cell body are seen now only isolated granules and threads, which barely stain
with eosin, while all other parts of the cell body appear entirely homogeneous,
transparent and colorless, till finally also the individual granules and threads
disappear, and the nucleus is now surrounded only by a perfectly transparent and
colorless cytoplasm. Whether the latter in the condition just described is to be
regarded as fluid, or has merely become extremely thin and transparent, we wish
to leave undecided. But as the change begins with a vacuolation and a break-
down of the protoplasm, and furthermore often the chromatin of the nucleus
becomes indistinct and finally a complete dissolution of the latter follows, we
have chosen the designation, liquefaction. In such islands, the remaining, not
yet liquefied epithelial cells show an exceedingly thin, delicate, but still not homo-
geneous and transparent protoplasm. It is therefore not improbable that a
thinning of the protoplasm precedes the above described vacuolation and
liquefaction."

These observations, beginning with 18 cases of diabetes, were ex-
tended so that in 1911 Weichselbaum was able to report 183 diabetic
necropsies, with "a, still larger number" of non-diabetic controls.
Of the diabetic cases, 53 per cent showed hydropic degeneration or
atrophy, 43 per cent fibrous, and 28 per cent hyaline changes. These
were described, as was inevitable under the circumstances and in the
absence of animal experiments, together as the three principal lesions
causing diabetes. The recognition of the hydropic change was facili-
tated by the fact that the diabetic material was mostly from severe
cases with termination in coma, and by the care taken for freshness

^^ For bibliography see Chapter I, also Allen, Glycosuria and Diabetes, Chap-
ter XXI.



ETIOLOGY AND PATHOLOGY 621

and fixation of tissue with a view to applying special stains. The
broad plan of the work also contributed notably to its success, a
careful study of the normal histology having been first made by
Stangl, later the development was studied in the fetus and infant
by Weichselbaum and Kyrle, and finally the pathology was closely
investigated in the largest series of diabetic and non-diabetic cases on
record. The admirable manner in which the research thus planned
was carried out makes it a classic. Both the descriptions and illus-
trations show clearly and convincingly the exact appearances en-
countered in the study of any long series of t)^ical cases of severe dia-
betes. The discussion is also worthy of such a foundation, the
necessity of examining many sections from different parts of the pan-
creas, the non-specificity of hemorrhages in the islands, the pictures
indicative of regeneration and hyperplasia, the consideration of quan-
titative changes, the answers to opponents of the insular hypothesis,
and the etiologic suggestions, being still worthy of notice.

The investigation carried on at Harvard, 1909 to 1912, developed an
improved form of experimental diabetes, in which hydropic degenera-
tion of islands was found to occur. Descriptions and photographs
were published in the former monograph.^^ The vacuolation was
proved to be specific to diabetes and to run parallel to the clinical
course, and was interpreted as exhaustion due to functional over^
strain. Professor F. B. Mallory, with whose aid and advice the
microscopic study was made, first suggested that the visible exhaus-
tion of cells might be the result of the diabetes.

Reference was made in that publication to the work of Lane and
Bensley, who proved by differential staining that the islands contain
two types of cells. What they called the a cells are few in num-
ber, while the fi cells make up the main mass of the island. The
different affinity of the fine (presumably secretory) granules of these
cells for dyes distinguishes them from each other and (along with
mitochondrial and other characteristics) also from acinar, duct,
centroacinar, a'nd imperfectly differentiated island cells.

Homans^" in 1912, using Bensley's stains, demonstrated the loss of
granules of the /3 cells, so that they came to resemble duct cells,

^^ AUen, Glycosuria and Diabetes, 1913.

^^Homans, -i^Proc. Roy.Soc. London, Series B, 1912, Ixxxvi, 73-87; /. Med.
Research, 1914, xxx, 49-68; 1915-16, xxxiii, 1-51.



622 CHAPTER VIII

in dogs made diabetic by removal of all the pancreas except a small
fragment at the tip of the uncinate process. He was inclined to in-
terpret it as a specific exhaustion due to functional overactivity, but
hesitated to regard the r61e of the islands in carbohydrate metabolism
as positively proved. In 1914 he pubhshed a fuller investigation,
altogether independent of that of Allen, showing that in cats mild
diabetes is accompanied by thinning and later complete loss of gran-
ules from the ;8 cells, while the nuclei and mitochondria remain
normal, and more severe diabetes is marked by degeneration and dis-
appearance especially of the fi cells, though a granules may
also remain in only a few cells. In 1915 he reported similar observa-
tions in dogs, indicating uniform preservation of a cells while ;8 cells
degenerated, the possible new formation of small islands composed
only of a cells, and the production of the hydropic process by carbo-
hydrate food, the sequence being defined as activity, exhaustion, and
degeneration. An illustration was also given of an exactly similar
vacuolation of the fi cells in a human diabetic, confirming Weichsel-
baimaand Stangl.

Hydropic degeneration has also been described in the above men-
tioned case of Ghon and Roman," and in an interesting and important
instance of spontaneous diabetes in a dog studied by Krumbhaar.^'

Martin,^* under the direction of H. M. Evans at Johns Hopkins,
tested various pure dyes of the group used by Bensley, and determined
that the best for the purpose is a combination of ethyl violet with
either orange G or azofuchsin. The latter may be particularly recom-
mended as an improvement upon Bensley's original stains, in that it is
easier and more reliable in apphcation and gives more vivid and dis-
tinct pictures. This work was continued in connection with the dia-
betic investigation in this Institute. The existence of the two kinds
of granules in normal island cells was confirmed. The work being
incomplete, Martin did not reach a conclusion whether the granules
were confined strictly to different cells. He suspected that there
might be gradations, and that two sorts of granules might be present
sometimes in the same cell, but owing to possible overlapping of cells
and other elements of doubt, he did not feel justified in opposing

" Krumbhaar, E. B., /. Exp. Med., 1916, xxiv, 361-365.
28 Martin, W. B., Anat. Rec., 191S, ix, 475^81.



ETIOLOGY AND PATHOLOGY 623

Bensley's carefully formed opinion that the two cell types are wholly
distinct. Romans' discovery that only the /3 cells are subject to the
characteristic exhaustion and degeneration was corroborated, and this
fact favors Bensley's view of the independence of the two types.
Often the a cells persist with if anything heavier than normal granu-
lation among the swollen and vacuolated p cells, as described by
Homans. This investigation was applied to dogs in which the metab-
ohsm and clinical condition had been carefully studied, in the hope
of gaining information of the function of the a cells, but no definite
relation between them and any phase of the diabetic disturbance was
observed. Even the persistence of a cells in dogs showing a "total"
D : N ratio could not prove them unconcerned in carbohydrate metab-
ohsm, since there is the possibility of a functional disabiHty, such as
must sometimes be assumed for the /3 cells, and also of functional
interrelations normally between the two types of cells. In the
extreme stages of experimental diabetes, vacuolation occurs in the
sniall ducts and ceU-cords, suggesting a real or attempted internal
secretory function in them, though this is uncertain.

The vital and' differential staining methods open important fields of
research. Knowledge of the normal and the comparative histology
of the pancreas is by no means complete. Though the essential in-
dependence of the islands is established by the work of Bensley and
by their specific degeneration in diabetes, the frequency and extent
of new formation of both islands and acini from ducts, the behavior
of the islands in inflammation, regeneration, etc., and the exact
border Hne between normal and pathologic processes, may be
mentioned as unsettled problems. In diabetes and other patho-
logic states, it is still sometimes impossible to distinguish between
acinar and island cells; the problem of regeneration and hyper-
plasia of island tissue has practical as well as theoretical bearings;
and the function and relations of the several types of island cells,
together with the duct and centroacinar cells, will be learned only by
stud3dng abnormal, as well as normal organs. A point of particular
interest was whether the special stains would furnish the long sought
infalHble anatomic diagnosis of diabetes. There was a chance that
the islands which cause confusion in some diabetic cases might prove
to be composed of a. cells only; also that sparseness of granulation



624 CHAPTER vni

might reveal diabetes in some instances where ordinary stains showed
normal appearances. These conditions may in fact occur to some
extent; but on the whole it is found, in agreement with conclusions
expressed by Homans in conversation, that the special stains do not
remove the difl&culties of diagnosis. They furthermore are rather
difi&cult and laborious in appKcation, they demand a freshness of
tissue and perfection of fixation seldom attainable in human autopsies,
and even under ideal conditions in pathologic material they often
fail to give the differentiation desired.

It is fortunate for practical convenience that routine methods care-
fully applied are sufficient. More or less can be learned from ordi-
nary pathologic specimens in Zenker or formaldehyde solution, and
neutral formaldehyde or Miiller-formol mixtures are preferred by
some. Instead of the usual 5 per cent acetic Zenker, the fixative of
choice in the present work has been either a plain solution of 2.5
per cent potassium bichromate and 5 per cent mercuric chloride, or
the same solution with addition of 1 or 2 per cent acetic acid just
before using. Lane and Bensley have found that stronger acidity
tends to dissolve out the specfic island granules. The most im-
portant consideration for the study of hydropic degeneration is the
freshness of the tissue. The autopsy should be performed immedi-
ately after death if possible, and the first step in it should be the re-
moval and weighing of the pancreas, followed by immediate fixation
of the pancreas specimens. It is advisable to take specimens from
different parts of the gland in separate bottles. The procedure here
has unconsciously imitated Weichselbaum and Stangl in taking them
in three sets, from the head, body, and tail respectively. It is also
desirable to take two kinds of specimens from each location: one,
pieces of ordinary size, for the purpose of examining the nimiber of
islands and the general pathology; the other, tiny bits of tissue, only
a few millimeters in dimensions for the sake of the quickest possible
penetration of the fixative for the study of the cytology of the islands
with either special or routine stains. For the latter the combina-
tion of eosin and either hematoxyhn or methylene blue has been satis-
factory in the present work. The fixation of other specimens in
formaldehyde for fat or in absolute alcohol for glycogen stains, and
other special measures, are of course added when desired. Persons



ETIOLOGY AND PATHOLOGY 625

contemplating an investigation, will do well to begin with animals, in
which the diabetic change is plainest, and then to examine some
human specimens showing it in typical advanced form, after which
they will be better equipped to judge more doubtful material.

Information concerning hydropic degeneration may be summarized
under six headings, as respects its description, nature, cause, mechan-
ism, consequences, and significance.

1. Description. — ^Histologically, the process begins with a thinning
of the fine granules, supposedly of internal secretion, shown in the
island cells by the Bensley methods, or a paling and clearing of the
finely granular cytoplasm seen with ordinary stains. The finest
routine preparations probably reveal the change almost as early and
delicately as the special dyes. The progress of the vacuolation or loss
of granules is as described by Weichselbaum and Stangl and by
Romans. The swelling of the individual cells, seemingly by imbibi-
tion of fluid, is apt to be more prominent in the dog and especially
the cat than in other species, such as man, monkey, and raccoon, but
occurs in all. In the most extreme degree, the islands composed of
swollen empty cells somewhat resemble adipose tissue; Homans com-
pares the network of cell membranes to a coarse meshed sieve; or the
exhausted islands may seem to blaze out against the dark back-
ground of the acinar tissue like snowballs or hydrangeas in full bloom.
Differences between species and between individual examples in
the same species are governed largely (though perhaps not solely)
by the acuteness of the process, giving both a more intense
swelling and a greater number of simultaneously exhausted cells
than when the change is slower. The alteration in the nucleus
seems to be strictly secondary. An apparently normal nucleus
can sometimes be seen naked, after the cell membrane has
burst. The nucleus does not, as might be implied by Weichselbaum
and Stangl's description, simply dissolve; but ordinarily at some,
generally an advanced, stage of vacuolation of the cytoplasm, it
becomes markedly dense and shrunken, and this pyknosis is one of
the prominent features of the process. The shrivelled nucleus then
dissolves, and is the last part of the cell to disappear. The observa-
tion of Weichselbaum and Stangl, that the hydropic change first begins
in occasional cells, while most cells in the same island and all cells in



626 CHAPTER vin

many islands still appear normal, holds for all species. Many /3 cells
remain free from vacuolation even when many others have com-
pletely disintegrated. In the more advanced stages, recourse to the
differential stains may be necessary to determine whether the re-
mainder are the usual a cells, or |8 cells persisting with slight alteration
or none. The order of precedence in which cells are affected is not
governed by their position at the center or periphery of the island,
or any obvious relation to the blood supply, or other known rules.
The loss of cells is not replaced by fibrous tissue; neither, except in
the most acute stage, are visible gaps left; but the acinar tissue
crowds in on all sides, whether by new formation of acini or other-
wise is unknown, and the reduced island remains as a compact group
of cells, until finally it may consist only of a barely distinguishable
clump of a cells and more or less framework, or may disappear with-
out a recognizable trace.

2. Nature. — There is nothing to oppose Weichselbaum and Stangl's
assumption that the vacuolation represents essentially water. The
ordinary appearances, in addition to Sudan stains, suffice to rule out
fat. Tests for protein and salts have not been made. Most observ-
ers are immediately impressed with the similarity to the Armanni or
Ehrlich change in certain cells of the renal tubules. Glycogen has
not been demonstrable by Best's carmine in the vacuolated islands.
The cause or nature of the renal picture calls for investigation. It has
been commonly accepted as "glycogenic infiltration" because of
EhrHch's demonstration of the presence of glycogen in diabetes.
The renal and pancreatic phenomena are not on a par, since the
former may be found in various conditions while the latter is specific
to diabetes. The truest resemblance is probably to the empty cells
found in the adrenal medulla when nervous or other stimuli exhaust
the fine granules which supposedly represent the precursor of epi-
nephrine stored in the chromaffin tissue. Bensley in conversation sug-
gested the possibihty that the Langerhans vacuoles may contain the
internal secretion in too great dilution for deposition in granules.

3. Cause. — Experimentally, the hydropic change can be shown to be
specific to diabetes. It does not follow pancreatic operations unless
diabetes is produced. Dogs made potentially diabetic by such opera-
tions can be kept even for years, and the preservation of their islands



ETIOLOGY AND PATHOLOGY 627

can be demonstrated not only microscopically but also by the un-



Online LibraryFrederick M. (Frederick Madison) AllenTotal dietary regulation in the treatment of diabetes → online text (page 65 of 76)